{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Fraud detection combining both Amazon Fraud Detector and Amazon SageMaker models in the same transaction"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Importing tools and libraries to be used"
]
},
{
"cell_type": "code",
"execution_count": 89,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"import io\n",
"import sys\n",
"import json\n",
"import uuid\n",
"import numpy as np \n",
"import pandas as pd\n",
"import matplotlib.pyplot as plt\n",
"import boto3\n",
"import time\n",
"import sklearn\n",
"import seaborn as sns\n",
"import sagemaker\n",
"\n",
"from IPython.display import clear_output\n",
"from datetime import datetime\n",
"from io import StringIO\n",
"from imblearn.under_sampling import RandomUnderSampler\n",
"from imblearn.over_sampling import SMOTENC\n",
"from collections import Counter\n",
"from sklearn.model_selection import train_test_split\n",
"from sklearn.datasets import dump_svmlight_file\n",
"from sklearn.metrics import balanced_accuracy_score, cohen_kappa_score, confusion_matrix\n",
"from sklearn.metrics import roc_curve, roc_auc_score, auc, roc_auc_score, classification_report\n",
"%matplotlib inline\n",
"from sagemaker import get_execution_role\n",
"from sagemaker.amazon.amazon_estimator import get_image_uri\n",
"from sagemaker.predictor import csv_serializer\n",
"from math import sqrt"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Investigate and process the data"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's start by downloading the dataset from: https://www.kaggle.com/mlg-ulb/creditcardfraud?select=creditcard.csv and upload it into the notebook file system as creditcard.csv."
]
},
{
"cell_type": "code",
"execution_count": 66,
"metadata": {},
"outputs": [],
"source": [
"# Resources and env variables setup\n",
"s3_resource = boto3.resource('s3')\n",
"afd_resource = boto3.client('frauddetector')\n",
"\n",
"# suffix is appended to detector and model name for uniqueness \n",
"sufx = datetime.now().strftime(\"%Y%m%d\")\n",
"# prefix is prepended to the sagamaker components\n",
"prefx = 'sagemaker-model'\n",
"# replace with the training bucket created in the CloudFormation\n",
"S3_BUCKET = \"afd-poc-trainingbucket-1i37svk9elcoe\"\n",
"# replace with the output bucket created in the CloudFormation\n",
"S3_OUT_BUCKET = \"afd-poc-outputbucket-1fhn20d7tumgg\"\n",
"# Replace the ARN Role with the resources created in CloudFormation stack\n",
"ARN_ROLE = \"arn:aws:iam::387461613214:role/afd-poc-NotebookInstanceExecutionRole-1FNQ41S8H2G68\" "
]
},
{
"cell_type": "code",
"execution_count": 45,
"metadata": {},
"outputs": [],
"source": [
"pd.set_option('display.max_rows', 500)\n",
"pd.set_option('display.max_columns', 500)\n",
"pd.set_option('display.width', 1000)\n",
"\n",
"data = pd.read_csv('creditcard.csv', delimiter=',')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's take a peek at our data (we only show a subset of the columns in the table):"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Index(['Time', 'V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class'], dtype='object')\n"
]
},
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" Time \n",
" V1 \n",
" V2 \n",
" V3 \n",
" V4 \n",
" V5 \n",
" V6 \n",
" V7 \n",
" V8 \n",
" V9 \n",
" V10 \n",
" V11 \n",
" V12 \n",
" V13 \n",
" V14 \n",
" V15 \n",
" V16 \n",
" V17 \n",
" V18 \n",
" V19 \n",
" V20 \n",
" V21 \n",
" V22 \n",
" V23 \n",
" V24 \n",
" V25 \n",
" V26 \n",
" V27 \n",
" V28 \n",
" Amount \n",
" Class \n",
" \n",
" \n",
" \n",
" \n",
" count \n",
" 284807.000000 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 2.848070e+05 \n",
" 284807.000000 \n",
" 284807.000000 \n",
" \n",
" \n",
" mean \n",
" 94813.859575 \n",
" 3.919560e-15 \n",
" 5.688174e-16 \n",
" -8.769071e-15 \n",
" 2.782312e-15 \n",
" -1.552563e-15 \n",
" 2.010663e-15 \n",
" -1.694249e-15 \n",
" -1.927028e-16 \n",
" -3.137024e-15 \n",
" 1.768627e-15 \n",
" 9.170318e-16 \n",
" -1.810658e-15 \n",
" 1.693438e-15 \n",
" 1.479045e-15 \n",
" 3.482336e-15 \n",
" 1.392007e-15 \n",
" -7.528491e-16 \n",
" 4.328772e-16 \n",
" 9.049732e-16 \n",
" 5.085503e-16 \n",
" 1.537294e-16 \n",
" 7.959909e-16 \n",
" 5.367590e-16 \n",
" 4.458112e-15 \n",
" 1.453003e-15 \n",
" 1.699104e-15 \n",
" -3.660161e-16 \n",
" -1.206049e-16 \n",
" 88.349619 \n",
" 0.001727 \n",
" \n",
" \n",
" std \n",
" 47488.145955 \n",
" 1.958696e+00 \n",
" 1.651309e+00 \n",
" 1.516255e+00 \n",
" 1.415869e+00 \n",
" 1.380247e+00 \n",
" 1.332271e+00 \n",
" 1.237094e+00 \n",
" 1.194353e+00 \n",
" 1.098632e+00 \n",
" 1.088850e+00 \n",
" 1.020713e+00 \n",
" 9.992014e-01 \n",
" 9.952742e-01 \n",
" 9.585956e-01 \n",
" 9.153160e-01 \n",
" 8.762529e-01 \n",
" 8.493371e-01 \n",
" 8.381762e-01 \n",
" 8.140405e-01 \n",
" 7.709250e-01 \n",
" 7.345240e-01 \n",
" 7.257016e-01 \n",
" 6.244603e-01 \n",
" 6.056471e-01 \n",
" 5.212781e-01 \n",
" 4.822270e-01 \n",
" 4.036325e-01 \n",
" 3.300833e-01 \n",
" 250.120109 \n",
" 0.041527 \n",
" \n",
" \n",
" min \n",
" 0.000000 \n",
" -5.640751e+01 \n",
" -7.271573e+01 \n",
" -4.832559e+01 \n",
" -5.683171e+00 \n",
" -1.137433e+02 \n",
" -2.616051e+01 \n",
" -4.355724e+01 \n",
" -7.321672e+01 \n",
" -1.343407e+01 \n",
" -2.458826e+01 \n",
" -4.797473e+00 \n",
" -1.868371e+01 \n",
" -5.791881e+00 \n",
" -1.921433e+01 \n",
" -4.498945e+00 \n",
" -1.412985e+01 \n",
" -2.516280e+01 \n",
" -9.498746e+00 \n",
" -7.213527e+00 \n",
" -5.449772e+01 \n",
" -3.483038e+01 \n",
" -1.093314e+01 \n",
" -4.480774e+01 \n",
" -2.836627e+00 \n",
" -1.029540e+01 \n",
" -2.604551e+00 \n",
" -2.256568e+01 \n",
" -1.543008e+01 \n",
" 0.000000 \n",
" 0.000000 \n",
" \n",
" \n",
" 25% \n",
" 54201.500000 \n",
" -9.203734e-01 \n",
" -5.985499e-01 \n",
" -8.903648e-01 \n",
" -8.486401e-01 \n",
" -6.915971e-01 \n",
" -7.682956e-01 \n",
" -5.540759e-01 \n",
" -2.086297e-01 \n",
" -6.430976e-01 \n",
" -5.354257e-01 \n",
" -7.624942e-01 \n",
" -4.055715e-01 \n",
" -6.485393e-01 \n",
" -4.255740e-01 \n",
" -5.828843e-01 \n",
" -4.680368e-01 \n",
" -4.837483e-01 \n",
" -4.988498e-01 \n",
" -4.562989e-01 \n",
" -2.117214e-01 \n",
" -2.283949e-01 \n",
" -5.423504e-01 \n",
" -1.618463e-01 \n",
" -3.545861e-01 \n",
" -3.171451e-01 \n",
" -3.269839e-01 \n",
" -7.083953e-02 \n",
" -5.295979e-02 \n",
" 5.600000 \n",
" 0.000000 \n",
" \n",
" \n",
" 50% \n",
" 84692.000000 \n",
" 1.810880e-02 \n",
" 6.548556e-02 \n",
" 1.798463e-01 \n",
" -1.984653e-02 \n",
" -5.433583e-02 \n",
" -2.741871e-01 \n",
" 4.010308e-02 \n",
" 2.235804e-02 \n",
" -5.142873e-02 \n",
" -9.291738e-02 \n",
" -3.275735e-02 \n",
" 1.400326e-01 \n",
" -1.356806e-02 \n",
" 5.060132e-02 \n",
" 4.807155e-02 \n",
" 6.641332e-02 \n",
" -6.567575e-02 \n",
" -3.636312e-03 \n",
" 3.734823e-03 \n",
" -6.248109e-02 \n",
" -2.945017e-02 \n",
" 6.781943e-03 \n",
" -1.119293e-02 \n",
" 4.097606e-02 \n",
" 1.659350e-02 \n",
" -5.213911e-02 \n",
" 1.342146e-03 \n",
" 1.124383e-02 \n",
" 22.000000 \n",
" 0.000000 \n",
" \n",
" \n",
" 75% \n",
" 139320.500000 \n",
" 1.315642e+00 \n",
" 8.037239e-01 \n",
" 1.027196e+00 \n",
" 7.433413e-01 \n",
" 6.119264e-01 \n",
" 3.985649e-01 \n",
" 5.704361e-01 \n",
" 3.273459e-01 \n",
" 5.971390e-01 \n",
" 4.539234e-01 \n",
" 7.395934e-01 \n",
" 6.182380e-01 \n",
" 6.625050e-01 \n",
" 4.931498e-01 \n",
" 6.488208e-01 \n",
" 5.232963e-01 \n",
" 3.996750e-01 \n",
" 5.008067e-01 \n",
" 4.589494e-01 \n",
" 1.330408e-01 \n",
" 1.863772e-01 \n",
" 5.285536e-01 \n",
" 1.476421e-01 \n",
" 4.395266e-01 \n",
" 3.507156e-01 \n",
" 2.409522e-01 \n",
" 9.104512e-02 \n",
" 7.827995e-02 \n",
" 77.165000 \n",
" 0.000000 \n",
" \n",
" \n",
" max \n",
" 172792.000000 \n",
" 2.454930e+00 \n",
" 2.205773e+01 \n",
" 9.382558e+00 \n",
" 1.687534e+01 \n",
" 3.480167e+01 \n",
" 7.330163e+01 \n",
" 1.205895e+02 \n",
" 2.000721e+01 \n",
" 1.559499e+01 \n",
" 2.374514e+01 \n",
" 1.201891e+01 \n",
" 7.848392e+00 \n",
" 7.126883e+00 \n",
" 1.052677e+01 \n",
" 8.877742e+00 \n",
" 1.731511e+01 \n",
" 9.253526e+00 \n",
" 5.041069e+00 \n",
" 5.591971e+00 \n",
" 3.942090e+01 \n",
" 2.720284e+01 \n",
" 1.050309e+01 \n",
" 2.252841e+01 \n",
" 4.584549e+00 \n",
" 7.519589e+00 \n",
" 3.517346e+00 \n",
" 3.161220e+01 \n",
" 3.384781e+01 \n",
" 25691.160000 \n",
" 1.000000 \n",
" \n",
" \n",
"
\n",
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"data.hist(bins=50,figsize=(20,15))\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Looks good, columns 𝑉𝑖 have been normalized to have 0 mean and unit standard deviation as the result of a PCA. Now, lets change the data to be Amazon Fraud Detector compatible."
]
},
{
"cell_type": "code",
"execution_count": 50,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Index(['time', 'va', 'vb', 'vc', 'vd', 've', 'vf', 'vg', 'vh', 'vi', 'vj', 'vk', 'vl', 'vm', 'vn', 'vo', 'vp', 'vq', 'vr', 'vs', 'vt', 'vu', 'vv', 'vw', 'vx', 'vy', 'vz', 'vaa', 'vab', 'amount', 'class'], dtype='object')\n"
]
}
],
"source": [
"# to lowercase\n",
"data.columns = map(str.lower, data.columns)\n",
"print(data.columns)"
]
},
{
"cell_type": "code",
"execution_count": 46,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Index(['Time', 'va', 'vb', 'vc', 'vd', 've', 'vf', 'vg', 'vh', 'vi', 'vj', 'vk', 'vl', 'vm', 'vn', 'vo', 'vp', 'vq', 'vr', 'vs', 'vt', 'vu', 'vv', 'vw', 'vx', 'vy', 'vz', 'vaa', 'vab', 'Amount', 'Class'], dtype='object')\n"
]
}
],
"source": [
"# mapping column names numbers to letters\n",
"\n",
"def standardize_headers(x):\n",
" if any(char.isdigit() for char in x):\n",
" if int(x[1:]) > 26:\n",
" return 'va'+chr(int(x[1:])+70)\n",
" return 'v'+chr(int(x[1:])+96)\n",
" return x\n",
"\n",
"data.rename(columns=standardize_headers, inplace=True)\n",
"print(data.columns)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Then change the timestamp and label column names"
]
},
{
"cell_type": "code",
"execution_count": 51,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" EVENT_LABEL \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 0.0 \n",
" -1.359807 \n",
" -0.072781 \n",
" 2.536347 \n",
" 1.378155 \n",
" -0.338321 \n",
" 0.462388 \n",
" 0.239599 \n",
" 0.098698 \n",
" 0.363787 \n",
" 0.090794 \n",
" -0.551600 \n",
" -0.617801 \n",
" -0.991390 \n",
" -0.311169 \n",
" 1.468177 \n",
" -0.470401 \n",
" 0.207971 \n",
" 0.025791 \n",
" 0.403993 \n",
" 0.251412 \n",
" -0.018307 \n",
" 0.277838 \n",
" -0.110474 \n",
" 0.066928 \n",
" 0.128539 \n",
" -0.189115 \n",
" 0.133558 \n",
" -0.021053 \n",
" 149.62 \n",
" 0 \n",
" \n",
" \n",
" 1 \n",
" 0.0 \n",
" 1.191857 \n",
" 0.266151 \n",
" 0.166480 \n",
" 0.448154 \n",
" 0.060018 \n",
" -0.082361 \n",
" -0.078803 \n",
" 0.085102 \n",
" -0.255425 \n",
" -0.166974 \n",
" 1.612727 \n",
" 1.065235 \n",
" 0.489095 \n",
" -0.143772 \n",
" 0.635558 \n",
" 0.463917 \n",
" -0.114805 \n",
" -0.183361 \n",
" -0.145783 \n",
" -0.069083 \n",
" -0.225775 \n",
" -0.638672 \n",
" 0.101288 \n",
" -0.339846 \n",
" 0.167170 \n",
" 0.125895 \n",
" -0.008983 \n",
" 0.014724 \n",
" 2.69 \n",
" 0 \n",
" \n",
" \n",
" 2 \n",
" 1.0 \n",
" -1.358354 \n",
" -1.340163 \n",
" 1.773209 \n",
" 0.379780 \n",
" -0.503198 \n",
" 1.800499 \n",
" 0.791461 \n",
" 0.247676 \n",
" -1.514654 \n",
" 0.207643 \n",
" 0.624501 \n",
" 0.066084 \n",
" 0.717293 \n",
" -0.165946 \n",
" 2.345865 \n",
" -2.890083 \n",
" 1.109969 \n",
" -0.121359 \n",
" -2.261857 \n",
" 0.524980 \n",
" 0.247998 \n",
" 0.771679 \n",
" 0.909412 \n",
" -0.689281 \n",
" -0.327642 \n",
" -0.139097 \n",
" -0.055353 \n",
" -0.059752 \n",
" 378.66 \n",
" 0 \n",
" \n",
" \n",
" 3 \n",
" 1.0 \n",
" -0.966272 \n",
" -0.185226 \n",
" 1.792993 \n",
" -0.863291 \n",
" -0.010309 \n",
" 1.247203 \n",
" 0.237609 \n",
" 0.377436 \n",
" -1.387024 \n",
" -0.054952 \n",
" -0.226487 \n",
" 0.178228 \n",
" 0.507757 \n",
" -0.287924 \n",
" -0.631418 \n",
" -1.059647 \n",
" -0.684093 \n",
" 1.965775 \n",
" -1.232622 \n",
" -0.208038 \n",
" -0.108300 \n",
" 0.005274 \n",
" -0.190321 \n",
" -1.175575 \n",
" 0.647376 \n",
" -0.221929 \n",
" 0.062723 \n",
" 0.061458 \n",
" 123.50 \n",
" 0 \n",
" \n",
" \n",
" 4 \n",
" 2.0 \n",
" -1.158233 \n",
" 0.877737 \n",
" 1.548718 \n",
" 0.403034 \n",
" -0.407193 \n",
" 0.095921 \n",
" 0.592941 \n",
" -0.270533 \n",
" 0.817739 \n",
" 0.753074 \n",
" -0.822843 \n",
" 0.538196 \n",
" 1.345852 \n",
" -1.119670 \n",
" 0.175121 \n",
" -0.451449 \n",
" -0.237033 \n",
" -0.038195 \n",
" 0.803487 \n",
" 0.408542 \n",
" -0.009431 \n",
" 0.798278 \n",
" -0.137458 \n",
" 0.141267 \n",
" -0.206010 \n",
" 0.502292 \n",
" 0.219422 \n",
" 0.215153 \n",
" 69.99 \n",
" 0 \n",
" \n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" EVENT_LABEL \n",
" EVENT_TIMESTAMP \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 0.0 \n",
" -1.359807 \n",
" -0.072781 \n",
" 2.536347 \n",
" 1.378155 \n",
" -0.338321 \n",
" 0.462388 \n",
" 0.239599 \n",
" 0.098698 \n",
" 0.363787 \n",
" 0.090794 \n",
" -0.551600 \n",
" -0.617801 \n",
" -0.991390 \n",
" -0.311169 \n",
" 1.468177 \n",
" -0.470401 \n",
" 0.207971 \n",
" 0.025791 \n",
" 0.403993 \n",
" 0.251412 \n",
" -0.018307 \n",
" 0.277838 \n",
" -0.110474 \n",
" 0.066928 \n",
" 0.128539 \n",
" -0.189115 \n",
" 0.133558 \n",
" -0.021053 \n",
" 149.62 \n",
" 0 \n",
" 2020-09-22T20:55:18Z \n",
" \n",
" \n",
" 1 \n",
" 0.0 \n",
" 1.191857 \n",
" 0.266151 \n",
" 0.166480 \n",
" 0.448154 \n",
" 0.060018 \n",
" -0.082361 \n",
" -0.078803 \n",
" 0.085102 \n",
" -0.255425 \n",
" -0.166974 \n",
" 1.612727 \n",
" 1.065235 \n",
" 0.489095 \n",
" -0.143772 \n",
" 0.635558 \n",
" 0.463917 \n",
" -0.114805 \n",
" -0.183361 \n",
" -0.145783 \n",
" -0.069083 \n",
" -0.225775 \n",
" -0.638672 \n",
" 0.101288 \n",
" -0.339846 \n",
" 0.167170 \n",
" 0.125895 \n",
" -0.008983 \n",
" 0.014724 \n",
" 2.69 \n",
" 0 \n",
" 2020-09-22T20:55:18Z \n",
" \n",
" \n",
" 2 \n",
" 1.0 \n",
" -1.358354 \n",
" -1.340163 \n",
" 1.773209 \n",
" 0.379780 \n",
" -0.503198 \n",
" 1.800499 \n",
" 0.791461 \n",
" 0.247676 \n",
" -1.514654 \n",
" 0.207643 \n",
" 0.624501 \n",
" 0.066084 \n",
" 0.717293 \n",
" -0.165946 \n",
" 2.345865 \n",
" -2.890083 \n",
" 1.109969 \n",
" -0.121359 \n",
" -2.261857 \n",
" 0.524980 \n",
" 0.247998 \n",
" 0.771679 \n",
" 0.909412 \n",
" -0.689281 \n",
" -0.327642 \n",
" -0.139097 \n",
" -0.055353 \n",
" -0.059752 \n",
" 378.66 \n",
" 0 \n",
" 2020-09-22T20:55:19Z \n",
" \n",
" \n",
" 3 \n",
" 1.0 \n",
" -0.966272 \n",
" -0.185226 \n",
" 1.792993 \n",
" -0.863291 \n",
" -0.010309 \n",
" 1.247203 \n",
" 0.237609 \n",
" 0.377436 \n",
" -1.387024 \n",
" -0.054952 \n",
" -0.226487 \n",
" 0.178228 \n",
" 0.507757 \n",
" -0.287924 \n",
" -0.631418 \n",
" -1.059647 \n",
" -0.684093 \n",
" 1.965775 \n",
" -1.232622 \n",
" -0.208038 \n",
" -0.108300 \n",
" 0.005274 \n",
" -0.190321 \n",
" -1.175575 \n",
" 0.647376 \n",
" -0.221929 \n",
" 0.062723 \n",
" 0.061458 \n",
" 123.50 \n",
" 0 \n",
" 2020-09-22T20:55:19Z \n",
" \n",
" \n",
" 4 \n",
" 2.0 \n",
" -1.158233 \n",
" 0.877737 \n",
" 1.548718 \n",
" 0.403034 \n",
" -0.407193 \n",
" 0.095921 \n",
" 0.592941 \n",
" -0.270533 \n",
" 0.817739 \n",
" 0.753074 \n",
" -0.822843 \n",
" 0.538196 \n",
" 1.345852 \n",
" -1.119670 \n",
" 0.175121 \n",
" -0.451449 \n",
" -0.237033 \n",
" -0.038195 \n",
" 0.803487 \n",
" 0.408542 \n",
" -0.009431 \n",
" 0.798278 \n",
" -0.137458 \n",
" 0.141267 \n",
" -0.206010 \n",
" 0.502292 \n",
" 0.219422 \n",
" 0.215153 \n",
" 69.99 \n",
" 0 \n",
" 2020-09-22T20:55:20Z \n",
" \n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" EVENT_LABEL \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 28515.0 \n",
" 1.226643 \n",
" 0.101988 \n",
" -0.087072 \n",
" 0.111524 \n",
" -0.281992 \n",
" -1.356027 \n",
" 0.469050 \n",
" -0.371725 \n",
" -0.153672 \n",
" -0.145105 \n",
" -0.143505 \n",
" 0.320964 \n",
" 0.149313 \n",
" 0.452515 \n",
" 0.776253 \n",
" 0.028739 \n",
" -0.177867 \n",
" -1.004746 \n",
" 0.264953 \n",
" 0.019692 \n",
" -0.355100 \n",
" -1.153663 \n",
" 0.109793 \n",
" 0.420318 \n",
" 0.197932 \n",
" 0.699218 \n",
" -0.114861 \n",
" 0.007583 \n",
" 50.40 \n",
" 0 \n",
" \n",
" \n",
" 1 \n",
" 83125.0 \n",
" 1.124848 \n",
" 0.125602 \n",
" 0.249962 \n",
" 0.489744 \n",
" -0.040386 \n",
" 0.167561 \n",
" -0.247614 \n",
" 0.284736 \n",
" -0.067302 \n",
" -0.170139 \n",
" 1.895117 \n",
" 0.507733 \n",
" -1.008648 \n",
" 0.187249 \n",
" 1.021663 \n",
" 0.136320 \n",
" 0.370186 \n",
" -0.573559 \n",
" -0.625413 \n",
" -0.204465 \n",
" -0.192467 \n",
" -0.576819 \n",
" 0.190343 \n",
" -0.357451 \n",
" 0.000870 \n",
" 0.139971 \n",
" -0.000993 \n",
" 0.011505 \n",
" 1.98 \n",
" 0 \n",
" \n",
" \n",
" 2 \n",
" 75537.0 \n",
" -0.307902 \n",
" 1.003715 \n",
" 1.404277 \n",
" 0.592627 \n",
" 0.311014 \n",
" -0.382106 \n",
" 0.531393 \n",
" -0.015292 \n",
" -0.758638 \n",
" -0.511597 \n",
" 0.149643 \n",
" 0.245866 \n",
" 0.752802 \n",
" -0.382214 \n",
" 1.590185 \n",
" -0.332546 \n",
" 0.611852 \n",
" -0.510495 \n",
" 0.563779 \n",
" 0.125220 \n",
" -0.131802 \n",
" -0.329268 \n",
" 0.046990 \n",
" 0.057413 \n",
" -0.656960 \n",
" 0.193192 \n",
" 0.142038 \n",
" 0.157501 \n",
" 1.98 \n",
" 0 \n",
" \n",
" \n",
" 3 \n",
" 156358.0 \n",
" 2.174919 \n",
" -1.535441 \n",
" -0.726428 \n",
" -1.430792 \n",
" -1.517258 \n",
" -0.751038 \n",
" -1.155344 \n",
" -0.180811 \n",
" -1.111885 \n",
" 1.536101 \n",
" -0.735705 \n",
" -0.771714 \n",
" 0.238603 \n",
" -0.447050 \n",
" -0.085886 \n",
" -0.568345 \n",
" 0.591162 \n",
" -0.104975 \n",
" -0.327292 \n",
" -0.334351 \n",
" -0.112766 \n",
" 0.050018 \n",
" 0.294666 \n",
" 1.123322 \n",
" -0.306025 \n",
" -0.241343 \n",
" 0.006553 \n",
" -0.027567 \n",
" 64.00 \n",
" 0 \n",
" \n",
" \n",
" 4 \n",
" 162523.0 \n",
" -2.221556 \n",
" 1.261987 \n",
" 2.047642 \n",
" 4.659268 \n",
" -0.535941 \n",
" 4.542044 \n",
" -3.715525 \n",
" -5.311701 \n",
" -0.955321 \n",
" 0.200601 \n",
" -1.342622 \n",
" 0.879905 \n",
" 0.241171 \n",
" -0.365540 \n",
" -1.735410 \n",
" 0.564495 \n",
" 0.380648 \n",
" 1.216692 \n",
" 1.872711 \n",
" 0.895990 \n",
" -1.820388 \n",
" 0.873723 \n",
" -2.648598 \n",
" -0.162180 \n",
" -0.492111 \n",
" 0.601490 \n",
" 0.627030 \n",
" 0.088289 \n",
" 379.29 \n",
" 0 \n",
" \n",
" \n",
" 5 \n",
" 113992.0 \n",
" -0.335198 \n",
" 0.871378 \n",
" 0.632703 \n",
" 4.164242 \n",
" 1.702582 \n",
" 1.954354 \n",
" 0.396722 \n",
" 0.495056 \n",
" -2.506020 \n",
" 1.609137 \n",
" 1.029249 \n",
" -0.143569 \n",
" -0.129187 \n",
" 0.862428 \n",
" 1.059507 \n",
" -0.862144 \n",
" 0.597711 \n",
" -0.509967 \n",
" 0.217672 \n",
" 0.260404 \n",
" 0.529635 \n",
" 1.473558 \n",
" 0.033413 \n",
" -1.333266 \n",
" -0.779961 \n",
" 0.595196 \n",
" 0.231547 \n",
" 0.193332 \n",
" 57.78 \n",
" 0 \n",
" \n",
" \n",
" 6 \n",
" 18653.0 \n",
" -1.305978 \n",
" 1.772118 \n",
" 0.741730 \n",
" 0.912351 \n",
" 0.498898 \n",
" 1.737490 \n",
" -0.957795 \n",
" -1.692301 \n",
" 0.755233 \n",
" -0.641461 \n",
" 0.900131 \n",
" -0.935381 \n",
" 3.068228 \n",
" 1.564764 \n",
" -1.357441 \n",
" -0.216156 \n",
" 0.712878 \n",
" 0.354701 \n",
" 0.959930 \n",
" -0.579386 \n",
" 1.985454 \n",
" -1.290109 \n",
" 0.108807 \n",
" -1.427937 \n",
" 0.140905 \n",
" -0.393444 \n",
" 0.078297 \n",
" -0.052505 \n",
" 1.00 \n",
" 0 \n",
" \n",
" \n",
" 7 \n",
" 133445.0 \n",
" -0.303356 \n",
" 1.144996 \n",
" -0.843639 \n",
" -1.046609 \n",
" 0.945826 \n",
" -1.822700 \n",
" 1.694510 \n",
" -0.465604 \n",
" -0.095435 \n",
" -0.150239 \n",
" -1.004530 \n",
" -0.509581 \n",
" -1.039596 \n",
" 0.805069 \n",
" -0.077089 \n",
" -0.649795 \n",
" -0.363229 \n",
" -0.307477 \n",
" -0.163633 \n",
" 0.007469 \n",
" 0.208148 \n",
" 0.842970 \n",
" -0.201490 \n",
" 0.076193 \n",
" -0.277275 \n",
" 0.093339 \n",
" 0.497711 \n",
" 0.336480 \n",
" 17.00 \n",
" 0 \n",
" \n",
" \n",
" 8 \n",
" 87252.0 \n",
" -0.465920 \n",
" 0.628365 \n",
" 1.449569 \n",
" 4.417351 \n",
" 1.034654 \n",
" 1.083905 \n",
" 0.103565 \n",
" -0.624355 \n",
" -0.226190 \n",
" 2.979203 \n",
" -1.319708 \n",
" -1.136836 \n",
" -0.027942 \n",
" -0.972251 \n",
" 0.758803 \n",
" -0.104912 \n",
" -0.462734 \n",
" 0.521972 \n",
" 1.049701 \n",
" -0.055464 \n",
" 0.192156 \n",
" 1.231734 \n",
" -0.113115 \n",
" 0.603061 \n",
" -1.890037 \n",
" 0.069854 \n",
" -0.907822 \n",
" -0.133170 \n",
" 11.31 \n",
" 0 \n",
" \n",
" \n",
" 9 \n",
" 40607.0 \n",
" 0.271095 \n",
" -2.720489 \n",
" 0.427427 \n",
" -0.080256 \n",
" -2.077311 \n",
" 0.334682 \n",
" -0.620412 \n",
" 0.174167 \n",
" 0.087480 \n",
" 0.347286 \n",
" 0.319218 \n",
" -0.528176 \n",
" -1.560458 \n",
" -0.281460 \n",
" -1.103553 \n",
" 1.181354 \n",
" 0.433565 \n",
" -0.623177 \n",
" 1.165483 \n",
" 1.068453 \n",
" 0.341998 \n",
" -0.188797 \n",
" -0.504412 \n",
" 0.027295 \n",
" 0.140939 \n",
" -0.294037 \n",
" -0.063847 \n",
" 0.102788 \n",
" 552.89 \n",
" 0 \n",
" \n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" EVENT_LABEL \n",
" EVENT_TIMESTAMP \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 28515.0 \n",
" 1.226643 \n",
" 0.101988 \n",
" -0.087072 \n",
" 0.111524 \n",
" -0.281992 \n",
" -1.356027 \n",
" 0.469050 \n",
" -0.371725 \n",
" -0.153672 \n",
" -0.145105 \n",
" -0.143505 \n",
" 0.320964 \n",
" 0.149313 \n",
" 0.452515 \n",
" 0.776253 \n",
" 0.028739 \n",
" -0.177867 \n",
" -1.004746 \n",
" 0.264953 \n",
" 0.019692 \n",
" -0.355100 \n",
" -1.153663 \n",
" 0.109793 \n",
" 0.420318 \n",
" 0.197932 \n",
" 0.699218 \n",
" -0.114861 \n",
" 0.007583 \n",
" 50.40 \n",
" 0 \n",
" 2020-09-23T05:01:16Z \n",
" \n",
" \n",
" 1 \n",
" 83125.0 \n",
" 1.124848 \n",
" 0.125602 \n",
" 0.249962 \n",
" 0.489744 \n",
" -0.040386 \n",
" 0.167561 \n",
" -0.247614 \n",
" 0.284736 \n",
" -0.067302 \n",
" -0.170139 \n",
" 1.895117 \n",
" 0.507733 \n",
" -1.008648 \n",
" 0.187249 \n",
" 1.021663 \n",
" 0.136320 \n",
" 0.370186 \n",
" -0.573559 \n",
" -0.625413 \n",
" -0.204465 \n",
" -0.192467 \n",
" -0.576819 \n",
" 0.190343 \n",
" -0.357451 \n",
" 0.000870 \n",
" 0.139971 \n",
" -0.000993 \n",
" 0.011505 \n",
" 1.98 \n",
" 0 \n",
" 2020-09-23T20:11:26Z \n",
" \n",
" \n",
" 2 \n",
" 75537.0 \n",
" -0.307902 \n",
" 1.003715 \n",
" 1.404277 \n",
" 0.592627 \n",
" 0.311014 \n",
" -0.382106 \n",
" 0.531393 \n",
" -0.015292 \n",
" -0.758638 \n",
" -0.511597 \n",
" 0.149643 \n",
" 0.245866 \n",
" 0.752802 \n",
" -0.382214 \n",
" 1.590185 \n",
" -0.332546 \n",
" 0.611852 \n",
" -0.510495 \n",
" 0.563779 \n",
" 0.125220 \n",
" -0.131802 \n",
" -0.329268 \n",
" 0.046990 \n",
" 0.057413 \n",
" -0.656960 \n",
" 0.193192 \n",
" 0.142038 \n",
" 0.157501 \n",
" 1.98 \n",
" 0 \n",
" 2020-09-23T18:04:58Z \n",
" \n",
" \n",
" 3 \n",
" 156358.0 \n",
" 2.174919 \n",
" -1.535441 \n",
" -0.726428 \n",
" -1.430792 \n",
" -1.517258 \n",
" -0.751038 \n",
" -1.155344 \n",
" -0.180811 \n",
" -1.111885 \n",
" 1.536101 \n",
" -0.735705 \n",
" -0.771714 \n",
" 0.238603 \n",
" -0.447050 \n",
" -0.085886 \n",
" -0.568345 \n",
" 0.591162 \n",
" -0.104975 \n",
" -0.327292 \n",
" -0.334351 \n",
" -0.112766 \n",
" 0.050018 \n",
" 0.294666 \n",
" 1.123322 \n",
" -0.306025 \n",
" -0.241343 \n",
" 0.006553 \n",
" -0.027567 \n",
" 64.00 \n",
" 0 \n",
" 2020-09-24T16:31:59Z \n",
" \n",
" \n",
" 4 \n",
" 162523.0 \n",
" -2.221556 \n",
" 1.261987 \n",
" 2.047642 \n",
" 4.659268 \n",
" -0.535941 \n",
" 4.542044 \n",
" -3.715525 \n",
" -5.311701 \n",
" -0.955321 \n",
" 0.200601 \n",
" -1.342622 \n",
" 0.879905 \n",
" 0.241171 \n",
" -0.365540 \n",
" -1.735410 \n",
" 0.564495 \n",
" 0.380648 \n",
" 1.216692 \n",
" 1.872711 \n",
" 0.895990 \n",
" -1.820388 \n",
" 0.873723 \n",
" -2.648598 \n",
" -0.162180 \n",
" -0.492111 \n",
" 0.601490 \n",
" 0.627030 \n",
" 0.088289 \n",
" 379.29 \n",
" 0 \n",
" 2020-09-24T18:14:44Z \n",
" \n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 41505.0 \n",
" -16.526507 \n",
" 8.584972 \n",
" -18.649853 \n",
" 9.505594 \n",
" -13.793819 \n",
" -2.832404 \n",
" -16.701694 \n",
" 7.517344 \n",
" -8.507059 \n",
" -14.110184 \n",
" 5.299236 \n",
" -10.834006 \n",
" 1.671120 \n",
" -9.373859 \n",
" 0.360806 \n",
" -9.899247 \n",
" -19.236292 \n",
" -8.398552 \n",
" 3.101735 \n",
" -1.514923 \n",
" 1.190739 \n",
" -1.127670 \n",
" -2.358579 \n",
" 0.673461 \n",
" -1.413700 \n",
" -0.462762 \n",
" -2.018575 \n",
" -1.042804 \n",
" 364.19 \n",
" \n",
" \n",
" 1 \n",
" 44261.0 \n",
" 0.339812 \n",
" -2.743745 \n",
" -0.134070 \n",
" -1.385729 \n",
" -1.451413 \n",
" 1.015887 \n",
" -0.524379 \n",
" 0.224060 \n",
" 0.899746 \n",
" -0.565012 \n",
" -0.087670 \n",
" 0.979427 \n",
" 0.076883 \n",
" -0.217884 \n",
" -0.136830 \n",
" -2.142892 \n",
" 0.126956 \n",
" 1.752662 \n",
" 0.432546 \n",
" 0.506044 \n",
" -0.213436 \n",
" -0.942525 \n",
" -0.526819 \n",
" -1.156992 \n",
" 0.311211 \n",
" -0.746647 \n",
" 0.040996 \n",
" 0.102038 \n",
" 520.12 \n",
" \n",
" \n",
" 2 \n",
" 35484.0 \n",
" 1.399590 \n",
" -0.590701 \n",
" 0.168619 \n",
" -1.029950 \n",
" -0.539806 \n",
" 0.040444 \n",
" -0.712567 \n",
" 0.002299 \n",
" -0.971747 \n",
" 0.756801 \n",
" 0.543827 \n",
" 0.112453 \n",
" 1.075384 \n",
" -0.245772 \n",
" 0.180483 \n",
" 1.769860 \n",
" -0.533172 \n",
" -0.533300 \n",
" 1.192245 \n",
" 0.212877 \n",
" 0.102398 \n",
" 0.168269 \n",
" -0.166639 \n",
" -0.810250 \n",
" 0.505083 \n",
" -0.232340 \n",
" 0.011409 \n",
" 0.004634 \n",
" 31.00 \n",
" \n",
" \n",
" 3 \n",
" 167123.0 \n",
" -0.432071 \n",
" 1.647895 \n",
" -1.669361 \n",
" -0.349504 \n",
" 0.785785 \n",
" -0.630647 \n",
" 0.276990 \n",
" 0.586025 \n",
" -0.484715 \n",
" -1.376648 \n",
" -1.328335 \n",
" 0.223621 \n",
" 1.132627 \n",
" -0.550875 \n",
" 0.616568 \n",
" 0.497974 \n",
" 0.502195 \n",
" 0.981343 \n",
" 0.101264 \n",
" -0.244633 \n",
" 0.358932 \n",
" 0.873663 \n",
" -0.178642 \n",
" -0.017171 \n",
" -0.207392 \n",
" -0.157756 \n",
" -0.237386 \n",
" 0.001934 \n",
" 1.50 \n",
" \n",
" \n",
" 4 \n",
" 168473.0 \n",
" 2.014160 \n",
" -0.137394 \n",
" -1.015839 \n",
" 0.327269 \n",
" -0.182179 \n",
" -0.956571 \n",
" 0.043241 \n",
" -0.160746 \n",
" 0.363241 \n",
" 0.259452 \n",
" 0.942162 \n",
" 0.850038 \n",
" -0.616166 \n",
" 0.592634 \n",
" -0.603845 \n",
" 0.091077 \n",
" -0.471867 \n",
" -0.333816 \n",
" 0.404711 \n",
" -0.255293 \n",
" -0.238644 \n",
" -0.616400 \n",
" 0.347045 \n",
" 0.061561 \n",
" -0.360196 \n",
" 0.174730 \n",
" -0.078043 \n",
" -0.070571 \n",
" 0.89 \n",
" \n",
" \n",
" 5 \n",
" 67878.0 \n",
" -0.641330 \n",
" -0.057304 \n",
" 1.489998 \n",
" -1.688131 \n",
" -1.151043 \n",
" 0.259996 \n",
" -1.391069 \n",
" -2.334075 \n",
" 1.168644 \n",
" -2.084080 \n",
" 0.480381 \n",
" 0.473738 \n",
" -2.192276 \n",
" 0.773942 \n",
" 0.294484 \n",
" 0.406074 \n",
" -0.541855 \n",
" 1.031450 \n",
" 0.017076 \n",
" 0.618411 \n",
" -1.231634 \n",
" 0.257164 \n",
" -0.371953 \n",
" -0.038566 \n",
" 1.397514 \n",
" -0.665947 \n",
" 0.031003 \n",
" 0.180357 \n",
" 100.00 \n",
" \n",
" \n",
" 6 \n",
" 159763.0 \n",
" 2.023952 \n",
" -0.120140 \n",
" -1.086918 \n",
" 0.423019 \n",
" -0.142901 \n",
" -1.127752 \n",
" 0.178493 \n",
" -0.303234 \n",
" 0.564509 \n",
" 0.062831 \n",
" -0.720047 \n",
" 0.366835 \n",
" -0.110857 \n",
" 0.319094 \n",
" 0.108359 \n",
" -0.153633 \n",
" -0.221312 \n",
" -0.934141 \n",
" 0.070553 \n",
" -0.210864 \n",
" -0.276175 \n",
" -0.697708 \n",
" 0.335631 \n",
" -0.017196 \n",
" -0.324904 \n",
" 0.200023 \n",
" -0.071566 \n",
" -0.058224 \n",
" 16.99 \n",
" \n",
" \n",
" 7 \n",
" 139631.0 \n",
" -0.688944 \n",
" 1.292153 \n",
" -0.564281 \n",
" -1.457526 \n",
" 2.258333 \n",
" -0.323270 \n",
" 1.678984 \n",
" -0.104128 \n",
" -1.285351 \n",
" -1.303435 \n",
" 0.282728 \n",
" -0.402525 \n",
" -0.548687 \n",
" -0.504283 \n",
" -0.685339 \n",
" 0.714828 \n",
" -0.092674 \n",
" 0.798953 \n",
" -0.150085 \n",
" -0.037150 \n",
" -0.006880 \n",
" -0.171568 \n",
" -0.720019 \n",
" -0.419435 \n",
" 1.211991 \n",
" 0.670916 \n",
" -0.103986 \n",
" 0.030084 \n",
" 8.95 \n",
" \n",
" \n",
" 8 \n",
" 133944.0 \n",
" 2.119362 \n",
" 0.142639 \n",
" -2.373337 \n",
" 0.541949 \n",
" 0.608419 \n",
" -1.775564 \n",
" 0.955775 \n",
" -0.599383 \n",
" 0.010420 \n",
" 0.295305 \n",
" -0.936569 \n",
" -0.452478 \n",
" -1.340798 \n",
" 1.077459 \n",
" -0.099584 \n",
" -0.815072 \n",
" 0.018481 \n",
" -0.639446 \n",
" -0.065427 \n",
" -0.323573 \n",
" 0.264264 \n",
" 0.898266 \n",
" -0.168063 \n",
" 0.059311 \n",
" 0.626949 \n",
" 0.729035 \n",
" -0.129120 \n",
" -0.094713 \n",
" 10.00 \n",
" \n",
" \n",
" 9 \n",
" 58769.0 \n",
" -5.584256 \n",
" -4.732413 \n",
" -0.448452 \n",
" -0.121442 \n",
" -0.707412 \n",
" -0.114376 \n",
" -1.554628 \n",
" 1.402126 \n",
" -0.031693 \n",
" -0.942358 \n",
" -2.439501 \n",
" -0.552312 \n",
" -0.295588 \n",
" -0.250246 \n",
" -1.197732 \n",
" 1.549553 \n",
" 0.933237 \n",
" -1.237689 \n",
" 0.416832 \n",
" -1.046900 \n",
" 0.041651 \n",
" 0.621789 \n",
" 0.223467 \n",
" -0.770137 \n",
" 0.621182 \n",
" -0.028738 \n",
" 0.505194 \n",
" -1.898323 \n",
" 101.49 \n",
" \n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" EVENT_LABEL \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 41505.0 \n",
" -16.526507 \n",
" 8.584972 \n",
" -18.649853 \n",
" 9.505594 \n",
" -13.793819 \n",
" -2.832404 \n",
" -16.701694 \n",
" 7.517344 \n",
" -8.507059 \n",
" -14.110184 \n",
" 5.299236 \n",
" -10.834006 \n",
" 1.671120 \n",
" -9.373859 \n",
" 0.360806 \n",
" -9.899247 \n",
" -19.236292 \n",
" -8.398552 \n",
" 3.101735 \n",
" -1.514923 \n",
" 1.190739 \n",
" -1.127670 \n",
" -2.358579 \n",
" 0.673461 \n",
" -1.413700 \n",
" -0.462762 \n",
" -2.018575 \n",
" -1.042804 \n",
" 364.19 \n",
" 1 \n",
" \n",
" \n",
" 1 \n",
" 44261.0 \n",
" 0.339812 \n",
" -2.743745 \n",
" -0.134070 \n",
" -1.385729 \n",
" -1.451413 \n",
" 1.015887 \n",
" -0.524379 \n",
" 0.224060 \n",
" 0.899746 \n",
" -0.565012 \n",
" -0.087670 \n",
" 0.979427 \n",
" 0.076883 \n",
" -0.217884 \n",
" -0.136830 \n",
" -2.142892 \n",
" 0.126956 \n",
" 1.752662 \n",
" 0.432546 \n",
" 0.506044 \n",
" -0.213436 \n",
" -0.942525 \n",
" -0.526819 \n",
" -1.156992 \n",
" 0.311211 \n",
" -0.746647 \n",
" 0.040996 \n",
" 0.102038 \n",
" 520.12 \n",
" 0 \n",
" \n",
" \n",
" 2 \n",
" 35484.0 \n",
" 1.399590 \n",
" -0.590701 \n",
" 0.168619 \n",
" -1.029950 \n",
" -0.539806 \n",
" 0.040444 \n",
" -0.712567 \n",
" 0.002299 \n",
" -0.971747 \n",
" 0.756801 \n",
" 0.543827 \n",
" 0.112453 \n",
" 1.075384 \n",
" -0.245772 \n",
" 0.180483 \n",
" 1.769860 \n",
" -0.533172 \n",
" -0.533300 \n",
" 1.192245 \n",
" 0.212877 \n",
" 0.102398 \n",
" 0.168269 \n",
" -0.166639 \n",
" -0.810250 \n",
" 0.505083 \n",
" -0.232340 \n",
" 0.011409 \n",
" 0.004634 \n",
" 31.00 \n",
" 0 \n",
" \n",
" \n",
" 3 \n",
" 167123.0 \n",
" -0.432071 \n",
" 1.647895 \n",
" -1.669361 \n",
" -0.349504 \n",
" 0.785785 \n",
" -0.630647 \n",
" 0.276990 \n",
" 0.586025 \n",
" -0.484715 \n",
" -1.376648 \n",
" -1.328335 \n",
" 0.223621 \n",
" 1.132627 \n",
" -0.550875 \n",
" 0.616568 \n",
" 0.497974 \n",
" 0.502195 \n",
" 0.981343 \n",
" 0.101264 \n",
" -0.244633 \n",
" 0.358932 \n",
" 0.873663 \n",
" -0.178642 \n",
" -0.017171 \n",
" -0.207392 \n",
" -0.157756 \n",
" -0.237386 \n",
" 0.001934 \n",
" 1.50 \n",
" 0 \n",
" \n",
" \n",
" 4 \n",
" 168473.0 \n",
" 2.014160 \n",
" -0.137394 \n",
" -1.015839 \n",
" 0.327269 \n",
" -0.182179 \n",
" -0.956571 \n",
" 0.043241 \n",
" -0.160746 \n",
" 0.363241 \n",
" 0.259452 \n",
" 0.942162 \n",
" 0.850038 \n",
" -0.616166 \n",
" 0.592634 \n",
" -0.603845 \n",
" 0.091077 \n",
" -0.471867 \n",
" -0.333816 \n",
" 0.404711 \n",
" -0.255293 \n",
" -0.238644 \n",
" -0.616400 \n",
" 0.347045 \n",
" 0.061561 \n",
" -0.360196 \n",
" 0.174730 \n",
" -0.078043 \n",
" -0.070571 \n",
" 0.89 \n",
" 0 \n",
" \n",
" \n",
"
\n",
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"#validating the training dataset\n",
"nonfrauds, frauds = data.groupby('EVENT_LABEL').size()\n",
"print('Number of frauds: ', frauds)\n",
"print('Number of non-frauds: ', nonfrauds)\n",
"print('Percentage of fradulent data:', 100.*frauds/(frauds + nonfrauds))\n",
"\n",
"count_class_0, count_class_1 = data.EVENT_LABEL.value_counts()\n",
"data.EVENT_LABEL.value_counts().plot(kind='bar', title='Count (EVENT_LABEL)');"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Uploading the data for training"
]
},
{
"cell_type": "code",
"execution_count": 62,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'ResponseMetadata': {'RequestId': 'AX5P2P5W6K2Y7V2W',\n",
" 'HostId': 'dAdIFZFsxxw5ufQqVlMhLaO1zn5bPeuI2w7ua3cqHKYcg8DcZ27gyL3OMuFDaOifAfwszauk39g=',\n",
" 'HTTPStatusCode': 200,\n",
" 'HTTPHeaders': {'x-amz-id-2': 'dAdIFZFsxxw5ufQqVlMhLaO1zn5bPeuI2w7ua3cqHKYcg8DcZ27gyL3OMuFDaOifAfwszauk39g=',\n",
" 'x-amz-request-id': 'AX5P2P5W6K2Y7V2W',\n",
" 'date': 'Tue, 22 Sep 2020 21:06:38 GMT',\n",
" 'x-amz-server-side-encryption': 'AES256',\n",
" 'etag': '\"8baed18fc24e97ddcba8e6ff15de1f93\"',\n",
" 'content-length': '0',\n",
" 'server': 'AmazonS3'},\n",
" 'RetryAttempts': 0},\n",
" 'ETag': '\"8baed18fc24e97ddcba8e6ff15de1f93\"',\n",
" 'ServerSideEncryption': 'AES256'}"
]
},
"execution_count": 62,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"csv_buffer = StringIO()\n",
"data.to_csv(csv_buffer, index=False)\n",
"s3_resource.Object(S3_BUCKET, 'dataset-training.csv').put(Body=csv_buffer.getvalue())"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Uploading the data for testing"
]
},
{
"cell_type": "code",
"execution_count": 63,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'ResponseMetadata': {'RequestId': 'B6E8DEF11A9A2C22',\n",
" 'HostId': 'jqD7xx2/tpv2PnFnhAiHXMRTaF7JDGex1nnP2GYXLFK/O19KkDONmxejes76LkFQtZE9mnbGlZU=',\n",
" 'HTTPStatusCode': 200,\n",
" 'HTTPHeaders': {'x-amz-id-2': 'jqD7xx2/tpv2PnFnhAiHXMRTaF7JDGex1nnP2GYXLFK/O19KkDONmxejes76LkFQtZE9mnbGlZU=',\n",
" 'x-amz-request-id': 'B6E8DEF11A9A2C22',\n",
" 'date': 'Tue, 22 Sep 2020 21:06:43 GMT',\n",
" 'x-amz-server-side-encryption': 'AES256',\n",
" 'etag': '\"4338feb9a18c65c7c0c98c68c5c935e6\"',\n",
" 'content-length': '0',\n",
" 'server': 'AmazonS3'},\n",
" 'RetryAttempts': 0},\n",
" 'ETag': '\"4338feb9a18c65c7c0c98c68c5c935e6\"',\n",
" 'ServerSideEncryption': 'AES256'}"
]
},
"execution_count": 63,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"csv_buffer = StringIO()\n",
"test.to_csv(csv_buffer, index=False)\n",
"\n",
"s3_resource.Object(S3_BUCKET, 'dataset-test.csv').put(Body=csv_buffer.getvalue())"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Once we have the datasets ready we need create the necesary entities for build and deploy the fraud detection model. This can be done within the Amazon Fraud Detector console or through the API as shown in the second jupyter notebook."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Next, we are going to traing and deploy the SageMaker version of the model."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"First we copy the data to an in-memory buffer."
]
},
{
"cell_type": "code",
"execution_count": 77,
"metadata": {},
"outputs": [],
"source": [
"buf = io.BytesIO()\n",
"\n",
"sklearn.datasets.dump_svmlight_file(X, y, buf)\n",
"buf.seek(0);"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now we upload the data to S3 using boto3."
]
},
{
"cell_type": "code",
"execution_count": 78,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Uploaded training data location: s3://afd-poc-trainingbucket-1i37svk9elcoe/sagemaker-dataset-training\n",
"Training artifacts will be uploaded to: s3://afd-poc-outputbucket-1fhn20d7tumgg\n"
]
}
],
"source": [
"key = 'sagemaker-dataset-training'\n",
"boto3.resource('s3').Bucket(S3_BUCKET).Object(key).upload_fileobj(buf)\n",
"\n",
"s3_train_data = 's3://{}/{}'.format(S3_BUCKET, key)\n",
"print('Uploaded training data location: {}'.format(s3_train_data))\n",
"\n",
"output_location = 's3://{}'.format(S3_OUT_BUCKET)\n",
"print('Training artifacts will be uploaded to: {}'.format(output_location))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can now train using SageMaker's built-in XGBoost algorithm. To specify the XGBoost algorithm, we use a utility function to obtain its URI. A complete list of built-in algorithms is found here: https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html"
]
},
{
"cell_type": "code",
"execution_count": 79,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"'get_image_uri' method will be deprecated in favor of 'ImageURIProvider' class in SageMaker Python SDK v2.\n"
]
}
],
"source": [
"container = get_image_uri(boto3.Session().region_name, 'xgboost', repo_version='1.0-1')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"SageMaker abstracts training via Estimators. We can pass the classifier and parameters along with hyperparameters to the estimator, and fit the estimator to the data in S3. An important parameter here is scale_pos_weight which scales the weights of the positive vs. negative class examples. This is crucial to do in an imbalanced dataset like the one we are using here, otherwise the majority class would dominate the learning."
]
},
{
"cell_type": "code",
"execution_count": 80,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Parameter image_name will be renamed to image_uri in SageMaker Python SDK v2.\n",
"'s3_input' class will be renamed to 'TrainingInput' in SageMaker Python SDK v2.\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"2020-09-22 21:27:11 Starting - Starting the training job...\n",
"2020-09-22 21:27:14 Starting - Launching requested ML instances......\n",
"2020-09-22 21:28:42 Starting - Preparing the instances for training............\n",
"2020-09-22 21:30:25 Downloading - Downloading input data...\n",
"2020-09-22 21:30:46 Training - Downloading the training image..\u001b[34mINFO:sagemaker-containers:Imported framework sagemaker_xgboost_container.training\u001b[0m\n",
"\u001b[34mINFO:sagemaker-containers:Failed to parse hyperparameter eval_metric value auc to Json.\u001b[0m\n",
"\u001b[34mReturning the value itself\u001b[0m\n",
"\u001b[34mINFO:sagemaker-containers:Failed to parse hyperparameter objective value binary:logistic to Json.\u001b[0m\n",
"\u001b[34mReturning the value itself\u001b[0m\n",
"\u001b[34mINFO:sagemaker-containers:No GPUs detected (normal if no gpus installed)\u001b[0m\n",
"\u001b[34mINFO:sagemaker_xgboost_container.training:Running XGBoost Sagemaker in algorithm mode\u001b[0m\n",
"\u001b[34m[21:31:19] 256326x30 matrix with 7688153 entries loaded from /opt/ml/input/data/train\u001b[0m\n",
"\u001b[34mINFO:root:Single node training.\u001b[0m\n",
"\u001b[34mINFO:root:Train matrix has 256326 rows\u001b[0m\n",
"\u001b[34m[0]#011train-auc:0.94930\u001b[0m\n",
"\u001b[34m[1]#011train-auc:0.97015\u001b[0m\n",
"\u001b[34m[2]#011train-auc:0.97634\u001b[0m\n",
"\u001b[34m[3]#011train-auc:0.97924\u001b[0m\n",
"\u001b[34m[4]#011train-auc:0.97912\u001b[0m\n",
"\u001b[34m[5]#011train-auc:0.98223\u001b[0m\n",
"\u001b[34m[6]#011train-auc:0.98227\u001b[0m\n",
"\u001b[34m[7]#011train-auc:0.98236\u001b[0m\n",
"\u001b[34m[8]#011train-auc:0.98251\u001b[0m\n",
"\u001b[34m[9]#011train-auc:0.98409\u001b[0m\n",
"\u001b[34m[10]#011train-auc:0.98539\u001b[0m\n",
"\u001b[34m[11]#011train-auc:0.98637\u001b[0m\n",
"\u001b[34m[12]#011train-auc:0.99252\u001b[0m\n",
"\u001b[34m[13]#011train-auc:0.99264\u001b[0m\n",
"\u001b[34m[14]#011train-auc:0.99328\u001b[0m\n",
"\u001b[34m[15]#011train-auc:0.99423\u001b[0m\n",
"\u001b[34m[16]#011train-auc:0.99657\u001b[0m\n",
"\u001b[34m[17]#011train-auc:0.99722\u001b[0m\n",
"\u001b[34m[18]#011train-auc:0.99788\u001b[0m\n",
"\n",
"2020-09-22 21:31:16 Training - Training image download completed. Training in progress.\u001b[34m[19]#011train-auc:0.99883\u001b[0m\n",
"\u001b[34m[20]#011train-auc:0.99899\u001b[0m\n",
"\u001b[34m[21]#011train-auc:0.99914\u001b[0m\n",
"\u001b[34m[22]#011train-auc:0.99910\u001b[0m\n",
"\u001b[34m[23]#011train-auc:0.99972\u001b[0m\n",
"\u001b[34m[24]#011train-auc:0.99973\u001b[0m\n",
"\u001b[34m[25]#011train-auc:0.99984\u001b[0m\n",
"\u001b[34m[26]#011train-auc:0.99989\u001b[0m\n",
"\u001b[34m[27]#011train-auc:0.99989\u001b[0m\n",
"\u001b[34m[28]#011train-auc:0.99990\u001b[0m\n",
"\u001b[34m[29]#011train-auc:0.99991\u001b[0m\n",
"\u001b[34m[30]#011train-auc:0.99993\u001b[0m\n",
"\u001b[34m[31]#011train-auc:0.99994\u001b[0m\n",
"\u001b[34m[32]#011train-auc:0.99995\u001b[0m\n",
"\u001b[34m[33]#011train-auc:0.99995\u001b[0m\n",
"\u001b[34m[34]#011train-auc:0.99996\u001b[0m\n",
"\u001b[34m[35]#011train-auc:0.99996\u001b[0m\n",
"\u001b[34m[36]#011train-auc:0.99996\u001b[0m\n",
"\u001b[34m[37]#011train-auc:0.99996\u001b[0m\n",
"\u001b[34m[38]#011train-auc:0.99996\u001b[0m\n",
"\u001b[34m[39]#011train-auc:0.99996\u001b[0m\n",
"\u001b[34m[40]#011train-auc:0.99997\u001b[0m\n",
"\u001b[34m[41]#011train-auc:0.99997\u001b[0m\n",
"\u001b[34m[42]#011train-auc:0.99997\u001b[0m\n",
"\u001b[34m[43]#011train-auc:0.99997\u001b[0m\n",
"\u001b[34m[44]#011train-auc:0.99997\u001b[0m\n",
"\u001b[34m[45]#011train-auc:0.99998\u001b[0m\n",
"\u001b[34m[46]#011train-auc:0.99998\u001b[0m\n",
"\u001b[34m[47]#011train-auc:0.99998\u001b[0m\n",
"\u001b[34m[48]#011train-auc:0.99998\u001b[0m\n",
"\u001b[34m[49]#011train-auc:0.99998\u001b[0m\n",
"\u001b[34m[50]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[51]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[52]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[53]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[54]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[55]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[56]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[57]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[58]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[59]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[60]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[61]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[62]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[63]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[64]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[65]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[66]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[67]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[68]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[69]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[70]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[71]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[72]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[73]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[74]#011train-auc:0.99999\u001b[0m\n",
"\u001b[34m[75]#011train-auc:0.99999\u001b[0m\n",
"\n",
"2020-09-22 21:32:14 Uploading - Uploading generated training model\u001b[34m[76]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[77]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[78]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[79]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[80]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[81]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[82]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[83]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[84]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[85]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[86]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[87]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[88]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[89]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[90]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[91]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[92]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[93]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[94]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[95]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[96]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[97]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[98]#011train-auc:1.00000\u001b[0m\n",
"\u001b[34m[99]#011train-auc:1.00000\u001b[0m\n",
"\n",
"2020-09-22 21:32:23 Completed - Training job completed\n",
"Training seconds: 118\n",
"Billable seconds: 118\n"
]
}
],
"source": [
"# Because the data set is so highly skewed, we set the scale position weight conservatively,\n",
"# as sqrt(num_nonfraud/num_fraud).\n",
"# Other recommendations for the scale_pos_weight are setting it to (num_nonfraud/num_fraud).\n",
"session = sagemaker.Session()\n",
"\n",
"scale_pos_weight = sqrt(np.count_nonzero(y==0)/np.count_nonzero(y))\n",
"hyperparams = {\n",
" \"max_depth\":5,\n",
" \"subsample\":0.8,\n",
" \"num_round\":100,\n",
" \"eta\":0.2,\n",
" \"gamma\":4,\n",
" \"min_child_weight\":6,\n",
" \"silent\":0,\n",
" \"objective\":'binary:logistic',\n",
" \"eval_metric\":'auc',\n",
" \"scale_pos_weight\": scale_pos_weight\n",
"}\n",
"\n",
"clf = sagemaker.estimator.Estimator(container,\n",
" get_execution_role(),\n",
" hyperparameters=hyperparams,\n",
" train_instance_count=1, \n",
" train_instance_type='ml.m4.xlarge',\n",
" output_path=output_location,\n",
" sagemaker_session=session)\n",
"clf.fit({'train': s3_train_data})"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now we deploy the estimator to and endpoint."
]
},
{
"cell_type": "code",
"execution_count": 82,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Parameter image will be renamed to image_uri in SageMaker Python SDK v2.\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"---------------!"
]
}
],
"source": [
"predictor = clf.deploy(initial_instance_count=1,\n",
" endpoint_name=\"fraud-detection-endpoint\",\n",
" instance_type='ml.m4.xlarge', \n",
" serializer=csv_serializer,\n",
" deserializer=None,\n",
" content_type='text/csv')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Once we have trained the model we can use it to make predictions for the test set."
]
},
{
"cell_type": "code",
"execution_count": 83,
"metadata": {},
"outputs": [],
"source": [
"# Because we have a large test set, we call predict on smaller batches\n",
"def predict(current_predictor, data, rows=500):\n",
" split_array = np.array_split(data, int(data.shape[0] / float(rows) + 1))\n",
" predictions = ''\n",
" for array in split_array:\n",
" predictions = ','.join([predictions, current_predictor.predict(array).decode('utf-8')])\n",
"\n",
" return np.fromstring(predictions[1:], sep=',')"
]
},
{
"cell_type": "code",
"execution_count": 84,
"metadata": {},
"outputs": [],
"source": [
"raw_preds = predict(predictor, X_test)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We will use a few measures from the scikit-learn package to evaluate the performance of our model. When dealing with an imbalanced dataset, we need to choose metrics that take into account the frequency of each class in the data.\n",
"Two such metrics are the balanced accuracy score, and Cohen's Kappa."
]
},
{
"cell_type": "code",
"execution_count": 201,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Balanced accuracy = 0.9021211611532022\n"
]
}
],
"source": [
"# sagamaker model balanced accuracy for 0.5 threshold\n",
"y_preds = np.where(raw_preds > 0.5, 1, 0)\n",
"print(\"Balanced accuracy = {}\".format(balanced_accuracy_score(y_test, y_preds)))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can already see that our model performs very well in terms of both metrics, Cohen's Kappa scores above 0.8 are generally very favorable.\n",
"Apart from single-value metrics, it's also useful to look at metrics that indicate performance per class. A confusion matrix, and per-class precision, recall and f1-score can also provide more information about the model's performance."
]
},
{
"cell_type": "code",
"execution_count": 202,
"metadata": {},
"outputs": [],
"source": [
"def plot_confusion_matrix(y_true, y_predicted):\n",
" cm = confusion_matrix(y_true, y_predicted)\n",
" # Get the per-class normalized value for each cell\n",
" cm_norm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
" # We color each cell according to its normalized value, annotate with exact counts.\n",
" ax = sns.heatmap(cm_norm, annot=cm, fmt=\"d\")\n",
" ax.set(xticklabels=[\"non-fraud\", \"fraud\"], yticklabels=[\"non-fraud\", \"fraud\"])\n",
" ax.set_ylim([0,2])\n",
" plt.title('Confusion Matrix')\n",
" plt.ylabel('Real Classes')\n",
" plt.xlabel('Predicted Classes')\n",
" plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 203,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAWsAAAEWCAYAAACg+rZnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3dd5xV1bn/8c+XoVkAFTBWEBX1qoka0cSCAQtqTLBGxV5yiSWxxZbEWHN/sURjcq0YO8ZYoglG7LHHApaoWAlKRIyCBRQEpjy/P/Ye7nGcOXNmOHvO2TPfd1775Tm7rLUOM3nOmmevtbYiAjMzq27dKt0AMzNrnYO1mVkOOFibmeWAg7WZWQ44WJuZ5YCDtZlZDjhY2xKTtJSkuyTNkXTbEpSzv6T7y9m2SpB0j6SDK90O61wcrLsQSftJmizpc0nvp0Fl6zIUvRfwNaB/RPygvYVExE0RMaoM7fkSSSMkhaQ7muzfKN3/SInlnClpfGvnRcTOEXF9O5tr1iwH6y5C0gnAxcD/Iwmsg4DLgF3LUPxg4M2IqCtDWVmZBWwpqX/BvoOBN8tVgRL+/5Rlwr9YXYCkfsDZwNERcUdEzIuI2oi4KyJOSs/pJeliSTPT7WJJvdJjIyTNkPRTSR+mvfJD02NnAacD+6Q99sOb9kAlrZH2YLun7w+RNE3SZ5LelrR/wf4nCq7bUtKkNL0ySdKWBccekXSOpCfTcu6XNKDIP8Mi4C/Avun1NcDewE1N/q1+J+ldSXMlPSdpeLp/J+DnBZ/znwXt+B9JTwLzgTXTfT9Mj18u6faC8s+T9JAklfwDNMPBuqvYAugN3FnknF8A3wY2BjYCNgdOKzi+EtAPWBU4HLhU0vIRcQZJb/2WiFg2Iq4u1hBJywC/B3aOiD7AlsCLzZy3AnB3em5/4CLg7iY94/2AQ4EVgZ7AicXqBm4ADkpf7whMAWY2OWcSyb/BCsAfgdsk9Y6Ie5t8zo0KrjkQGAv0AaY3Ke+nwDfSL6LhJP92B4fXebA2crDuGvoDs1tJU+wPnB0RH0bELOAskiDUqDY9XhsRE4HPgXXb2Z4GYENJS0XE+xExpZlzdgHeiogbI6IuIm4GXge+X3DOtRHxZkR8AdxKEmRbFBH/AFaQtC5J0L6hmXPGR8RHaZ0XAr1o/XNeFxFT0mtqm5Q3HziA5MtmPPCTiJjRSnlmX+Fg3TV8BAxoTEO0YBW+3Cucnu5bXEaTYD8fWLatDYmIecA+wBHA+5LulrReCe1pbNOqBe//04723Aj8GBhJM39ppKme19LUy6ckf00US68AvFvsYEQ8C0wDRPKlYtZmDtZdw1PAAmC3IufMJLlR2GgQX00RlGoesHTB+5UKD0bEfRGxA7AySW/5qhLa09im99rZpkY3AkcBE9Ne72JpmuIUklz28hGxHDCHJMgCtJS6KJrSkHQ0SQ99JnBy+5tuXZmDdRcQEXNIbgJeKmk3SUtL6iFpZ0nnp6fdDJwmaWB6o+50kj/b2+NFYBtJg9Kbmz9rPCDpa5JGp7nrhSTplPpmypgIrJMON+wuaR9gfeBv7WwTABHxNvAdkhx9U32AOpKRI90lnQ70LTj+AbBGW0Z8SFoH+BVJKuRA4GRJRdM1Zs1xsO4iIuIi4ASSm4azSP50/zHJCAlIAspk4CXgZeD5dF976noAuCUt6zm+HGC7kdx0mwl8TBI4j2qmjI+A76XnfkTSI/1eRMxuT5ualP1ERDT3V8N9wD0kw/mmk/w1UpjiaJzw85Gk51urJ007jQfOi4h/RsRbJCNKbmwcaWNWKvmmtJlZ9XPP2swsBxyszcxywMHazCwHHKzNzHKg2CSJivreoF1859O+4razNqh0E6wKLXXo+Uu81krt7Gklx5weA9bs8LVd3LM2M8uBqu1Zm5l1qIbm5mZVDwdrMzOA+mpejt3B2swMgIiGSjehKAdrMzOABgdrM7Pq5561mVkO+AajmVkOuGdtZlb9wqNBzMxywDcYzcxywGkQM7Mc8A1GM7MccM/azCwHfIPRzCwHfIPRzKz6RThnbWZW/ZyzNjPLAadBzMxywD1rM7McqK+tdAuKcrA2MwOnQczMcsFpEDOzHHDP2swsBxyszcyqX/gGo5lZDjhnbWaWA06DmJnlgHvWZmY54J61mVkOuGdtZpYDdX74gJlZ9XPP2swsB5yzNjPLAfeszcxywD1rM7MccM/azCwHPBrEzCwHIirdgqIcrM3MwDlrM7NcqPJg3a3SDTAzqwrRUPrWCkk7SXpD0lRJpzZzfJCkhyW9IOklSd9trUz3rM3MAOrry1KMpBrgUmAHYAYwSdKEiHi14LTTgFsj4nJJ6wMTgTWKletgbWYG5UyDbA5MjYhpAJL+BOwKFAbrAPqmr/sBM1sr1MHazAzaFKwljQXGFuwaFxHj0terAu8WHJsBfKtJEWcC90v6CbAMsH1rdTpYm5lBmybFpIF5XAuH1dwlTd6PAa6LiAslbQHcKGnDiJYb4WBtZgZEQ9nGWc8AVi94vxpfTXMcDuwEEBFPSeoNDAA+bKlQjwYxM4MkDVLqVtwkYKikIZJ6AvsCE5qc829gOwBJ/wX0BmYVK9Q9azMzKNtokIiok/Rj4D6gBrgmIqZIOhuYHBETgJ8CV0k6niRFckhE8SmUDtZmZlDWSTERMZFkOF7hvtMLXr8KbNWWMh2szcyg6mcwOlhXudGHjWbHMTuCxH0338eEq/9a6SZZBSysq+ewm56gtq6Bugi2X3cVjhq+HoeOf5x5i5LV4j6Zv5ANVl6ei/dsOkrMSuKFnKy9Bq8zmB3H7MgJ3z+B2tpazr7xHCY/NImZ77Q6ft46mZ413bhqzFYs3bM7tfUNHDr+cbZec0WuPWD44nN+esezjBi6UgVbmXNV3rP2aJAqttrQ1Xn9+TdYuGAhDfUNvPL0y2yx0xaVbpZVgCSW7pn0reoaGqhrCFQwmnfewlqenT6bkeusXKEWdgINUfpWAWXvWUu6i68OAF8sIkaXu87Oavob0znopIPos1wfFi1YxLCRw3jrpamVbpZVSH1DMOa6R3j3k3ns880hfH2VFRYf+/ub7/OtNQawbK8eFWxhzpVpNEhWskiD/Cb97x7ASsD49P0Y4J1iFxZO4fz68hsyaNlBGTQvP2ZMfZfbL7+dc276FQvmL+Dt196mvsp/oSw7Nd3ErYeNZO6CWk6441mmzprL2gOT5SXufe09dv/G4Aq3MN+iytMgZQ/WEfEogKRzImKbgkN3SXqslWsXT+H83qBdqjvb30EeuOV+HrjlfgAOOvkgZr//UYVbZJXWt3cPhg3qz5PTPmTtgX359ItFvDLzEy7aY/NKNy3fKpTeKFWWOeuBktZsfCNpCDAww/o6pX79+wEwcJWBbLHTljw64dEKt8gq4eP5C5m7oBaABbX1PPPOLIb0XxaAB15/j+Frr0Sv7jWVbGL+lXE96yxkORrkeOARSdPS92sAP8qwvk7p51f+nD7L96W+to4rfnk58+Z8XukmWQXM/nwBv/zbCzRE0BDBqPVWZZu1k5Ef9776HodtMbTCLewEqrxnnVmwjoh7JQ0F1kt3vR4RC7Oqr7M6Za9TKt0EqwLrrNiPWw4b0eyxq/ffukPb0mnVVff9oMyCtaSDmuzaSBIRcUNWdZqZtVuF0hulyjINslnB694kK0w9DzhYm1n16cJpkJ8UvpfUD7gxq/rMzJZElxu6V8R8wHdBzKw6ddWedZOZjN2A9YFbs6rPzGyJdNVgzf/NZASoA6ZHxIwM6zMza78qnx2cZc7aszfMLDfK+AzGTGQ2g1HStyVNkvS5pEWS6iXNzao+M7Ml0tVW3StwCcmDIm8DhgEHAWtnWJ+ZWft15dEgETFVUk1E1APXSvpHlvWZmbVbladBsgzW89PHsL8o6XzgfWCZDOszM2u/Kg/WWa66d2Ba/o+BecDqwJ4Z1mdm1m5R31DyVgmZ9Kwl1QD/ExEHAAuAs7Kox8ysbKq8Z51JsI6IekkDJfWMiEVZ1GFmVk7VPnQvy5z1O8CTkiaQpEEAiIiLMqzTzKx9qjxYlz1nLalxsaZ9gL+ldfQp2MzMqk9DG7YKyKJnvamkwcC/gf/NoHwzs7KLuq43zvoK4F5gCDC5YL9IFnZas7mLzMwqqrpjdSZPN/898HtJl0fEkeUu38wsC9V+g7HVnLWkZSR1S1+vI2m0pB6tXedAbWa5UuU561JuMD4G9Ja0KvAQcChwXZaNMjPraNEQJW+VUEqwVkTMB/YA/jcidid5kICZWedR5T3rUnLWkrQFsD9weBuuMzPLjairdAuKKyXoHgf8DLgzIqZIWhN4ONtmmZl1rKjy0SCtpkEi4tGIGE2yPjURMS0ijsm8ZWZmHamMaRBJO0l6Q9JUSae2cM7ekl6VNEXSH1srs5TRIFtIehV4LX2/kaTLWm+umVl+REPpWzHpQnaXAjuT3N8bI2n9JucMJclYbBURG5BkMIoq5QbjxcCOwEcAEfFPYJsSrjMzy41yBWtgc2BqmoVYBPwJ2LXJOf8NXBoRnwBExIetFVrS2iAR8W6TXdX9GGAzszaKepW8SRoraXLBNragqFWBwpg5I91XaB1gHUlPSnpa0k6tta+UG4zvStoSiPTJL8eQpkTMzDqLttxgjIhxwLgWDqu5S5q87w4MBUYAqwGPS9owIj5tqc5SetZHAEeTfDPMADZO35uZdRrRoJK3VswgeTJWo9WAmc2c89eIqI2It4E3SIJ3i1rtWUfEbJIx1mZmnVYZh+5NAoZKGgK8B+wL7NfknL8AY4DrJA0gSYtMK1ZoKaNBzpfUV1IPSQ9Jmi3pgHZ9BDOzKhWhkrfi5UQdybNn7yNJGd+azlE5W9Lo9LT7gI/SkXYPAydFxEfFyi0lZz0qIk6WtDtJ1/0HaeHjS7jWzCwXyjkpJiImAhOb7Du94HUAJ6RbSUoJ1o0r7H0XuDkiPpZazdmYmeVKQ311x7VSgvVdkl4HvgCOkjSQ5InlZmadRgk3DiuqlBuMp0o6D5ibPrV8Hl8d4G1mlmvVHqxLucH4A6AuDdSnkeSqV8m8ZWZmHSii9K0SShln/cuI+EzS1iTTzq8HLs+2WWZmHauM46wzUUqwbpxavgtweUT8FeiZXZPMzDpeuYbuZaWUG4zvSboS2B44T1IvSlxTxMwsL+qrfDRIKUF3b5IB3Dul89ZXAE7KtFVmZh0s9z3r9PmLd0haUdKgdPfr2TbLzKxjdYbRIKMlvQW8DTya/veerBtmZtaROsNokHOAbwNvRsQQktz1k5m2ysysg3WG0SC16QIj3SR1i4iHSZZJNTPrNOobupW8VUIpo0E+lbQs8Bhwk6QPgSp/aLuZWdtUKr1RqlK+InYlWRfkeOBe4F/A97NslJlZR2sIlbxVQimjQeYVvL0+w7aYmVVMpYbklarFYC3pM7763DBIni8WEdE3s1aZmXWwak+DtBisI6JPRzbEzKySKpXeKJWiha8TSZsBAyLinib7vw/MjIjnsmxY7expVf49Z5Ww1CrDK90Eq0J1i95b4kj7zCp7lBxzvjXzjg6P7MVuMF5A8vywpl5Lj5mZdRrRhq0Sit1g7B8R7zTdGRFTJfXPrklmZh2v2tMgxYL1UkWOLVPuhpiZVVK1jwYplgZ5UNL/qMnTcSWdBfw922aZmXWshjZslVCsZ/1T4A/AVEkvpvs2AiYDP8y6YWZmHSmo7p51saF784AxktYENkh3T4mIaR3SMjOzDlRX5WmQUmYwTgMcoM2sU8ttz9rMrCupVC66VA7WZmbkuGctaYViF0bEx+VvjplZZeS5Z/0cyWSd5r5uAlgzkxaZmVVAfV571ukjvMzMuoQqf15uaTlrScsDQ4Hejfsi4rGsGmVm1tEa8tqzbiTph8CxwGrAiyQPz30K2DbbppmZdZxqX+azlMd6HQtsBkyPiJHAJsCsTFtlZtbB8jzdvNGCiFggCUm9IuJ1Setm3jIzsw7UoJynQYAZkpYD/gI8IOkTYGa2zTIz61j1lW5AK1pNg0TE7hHxaUScCfwSuBrYLeuGmZl1pAaVvrVG0k6S3pA0VdKpRc7bS1JIGtZamaWOBtkaGBoR10oaCKwKvF3KtWZmeVCu0SCSaoBLgR2AGcAkSRMi4tUm5/UBjgGeKaXcVnvWks4ATgF+lu7qAYwvvelmZtWvjI/12hyYGhHTImIR8Cdg12bOOwc4H1hQSvtKGQ2yOzAamAcQETMBP/nczDqVtqRBJI2VNLlgG1tQ1KrAuwXvZ6T7FpO0CbB6RPyt1PaVkgZZFBEhKdJK/EgvM+t02jIkLyLGAeNaONzSEh3JQakb8FvgkDZUWVLP+lZJVwLLSfpv4EGSJ8iYmXUa9Sp9a8UMYPWC96vx5RF0fYANgUckvUMy0XBCazcZS3n4wG8k7QDMBdYFTo+IB1ptrplZjpRxssskYKikIcB7wL7Afo0HI2IOMKDxvaRHgBMjYnKxQksaDZIG5wfSgmsk7R8RN7X1E5iZVatyBeuIqJP0Y+A+oAa4JiKmSDobmBwRE9pTbrH1rPsCR5MkxieQBOujgZNI1ghxsDazTqOcj2CMiInAxCb7Tm/h3BGllFmsZ30j8AnJok0/JAnSPYFdI+LFIteZmeVOnh8+sGZEfB1A0h+A2cCgiPisQ1pmZtaBqn26ebFgXdv4IiLqJb3tQG1mnVWeHz6wkaS56WsBS6XvBURE9M28dWZmHSS3aZCIqOnIhpiZVVJug7WZWVdS7U+KcbA2MyPfOWszsy4jz6NBzMy6jIYqT4Q4WJuZ4RuMZma5UN39agdrMzPAPWszs1yoU3X3rR2szcxwGsTMLBecBjEzywEP3TMzy4HqDtUO1mZmgNMgZma5UF/lfWsHazMz3LM2M8uFcM/azKz6uWdtALz/wSx+fs5vmP3xJ3ST2GvXnTlw7914/c1/cfYF/8vCRbXU1NTwyxOP5uvrr7v4updfe4P9x57Ab84+lVEjhzPzPx9w3M9/RX19A3V1dey312j22X0XvliwgBNO+3/MeO99unXrxoitv8XxRx5WwU9sWenVqxeP/P3P9OzVi+7da7jjjrs56+wLK92s3PPQPQOge00NJ/3kv1l/3bWZN28+ex9+DFtutgkXXnY1Rx62P8O32IzH/vEsF152Ndddcj4A9fX1/Paya9lq828uLmdg/xUYf8WF9OzZk/nzv2C3A49g5Nbfpk+fZTh0zJ5svulG1NbWcvgxP+PxpyYxfIvNKvWRLSMLFy5k+1F7M2/efLp3785jj9zJvfc+zDPPPl/ppuVadYdqB+sOM3DACgwcsAIAyyyzNGsOXp0PZn2EJD6fNx+Az+fNZ8UB/Rdf88fbJ7DDiK145bU3F+/r0aPH4teLamtpiORXbKnevdl8040Wn/Nf667NB7NmZ/65rDLmpb8zPXp0p3uPHkRUe6ipfnVVHq67ZVGopBWKbVnUmSfvvf8Br731L76xwbqccuyPuPCyq9lu9wP5zSV/4LgjDgHgg1mzeeixf7D3bt/9yvXvfzCL3Q86ku13P4jD9/8BKw7s/6Xjcz/7nEeffIZvbbpxR3wcq4Bu3boxedL9vP/eSzz00GM8O+mFSjcp96IN/6uETII18BwwOf3vLOBN4K309XMtXSRprKTJkib/4YabM2paZc2f/wXH/+JXnHLMj1h2mWW45c67OeUnY3nozhs5+ZixnP7riwE473dXcvyRh1FT89WHzK/8tYHcecPlTLzlav56z4PM/viTxcfq6uo5+czz2H+v0ay+6sod9rmsYzU0NDBss1EMHjKMzYZtwgYbrNv6RVZUQxu2SsgkDRIRQwAkXQFMiIiJ6fudge2LXDcOGAdQO3tadf9N0g61dXUc94tfscuokewwYisAJtzzID877ggAdtx2OGecmwTrKa+/xUlnnAvAJ3Pm8vhTk6ipqWG7bbZcXN6KA/uz9pDBPP/PVxg1cjgAZ57/OwattgoH7rN7R340q5A5c+by6GP/YMdRI5gy5Y1KNyfXqn3oXlY960abNQZqgIi4B/hOxnVWpYjg9F9fzJqDV+fgffdYvH/ggP5MeuFlAJ557kUGr74qAPfdfh33//l67v/z9YwasTWnnXg0222zJf/5cBYLFi4EYM7cz3jh5VdZY9BqAPx+3PV8/vl8Tj32Rx386awjDRiwAv369QWgd+/ebLftcN54418VblX+dcmedYHZkk4DxpPcbD0A+CjjOqvSCy9N4a57H2LoWmuw58FHA3Dsjw7mrFOO4dzfXUldfT29evbkjJOPKVrOtHfe5YJLrkISEcEhY/ZgnbWG8J8PZzHu+j8xZPDq/ODQnwAwZs/vs9fonTL/bNaxVl75a1xz9cXU1HSjW7du3H77Xdw98cFKNyv36qv8Jq2yvIuc3kw8A9gm3fUYcFZEfNzatZ0xDWJLbqlVhle6CVaF6ha9pyUtY7/Bu5ccc/44/c4lrq+tMu1Zp0H52CzrMDMrh2rPWWcarCU9TDNjzSNi2yzrNTNrq64+3fzEgte9gT2BuozrNDNrsy493Twimo6pflLSo1nWaWbWHuVMg0jaCfgdUAP8ISLObXL8BOCHJJ3XWcBhETG9WJlZp0EKZyt2AzYFVsqyTjOz9ijXaBBJNcClwA7ADGCSpAkR8WrBaS8AwyJivqQjgfOBfYqVm3Ua5DmSnLVIvkHeBg7PuE4zszYrYxpkc2BqREwDkPQnYFdgcbCOiIcLzn+aZFhzUVmnQYZkWb6ZWbm05QajpLHA2IJd49IZ2ACrAu8WHJsBfKtIcYcD97RWZ+ar7knaEFif5AYjABFxQ9b1mpm1RVty1oVLYzSjuTHYzRYu6QBgGCXM7M46Z30GMIIkWE8EdgaeAByszayqlDENMgNYveD9asDMpidJ2h74BfCdiFjYWqFZrw2yF7Ad8J+IOBTYCOiVcZ1mZm0WESVvrZgEDJU0RFJPYF9gQuEJkjYBrgRGR8SHpbQv6zTIFxHRIKlOUl/gQ2DNjOs0M2uz+jL1rCOiTtKPgftIhu5dExFTJJ0NTI6ICcAFwLLAbZIA/h0Ro4uVm3WwnixpOeAqkpEhnwPPZlynmVmblXNSTLra6MQm+04veN3iUtEtySxYK/m6+HVEfApcIeleoG9EvJRVnWZm7VXtj0bLLFhHREj6C8lEGCLinazqMjNbUtU+3TzrG4xPS/Ljtc2s6lX7MxizzlmPBH4kaTowj2T8YUTENzKu18ysTar94QOZBGtJQyLibZJx1WZmVa/a0yBZ9axvJ8lVXxMR22VUh5lZ2XTVYN0tnb24TroU4JdExEUZ1Wtm1i5ddTTIvsBuafl9MqrDzKxsumTPOiLeAM6T9FJEtLqalJlZpVX7MxgzHbpXGKgl/S3LuszMlkR9NJS8VULmS6QWWLUD6zIza5OumrNuzgsdWJeZWZt0yZx1cyLisI6qy8ysrao9Z531wwe2As4EBqd1Nc5g9DKpZlZVGrp4GuRq4HiS5VHrM67LzKzdunTPGpjjoXtmlgeVGuVRqqyD9cOSLgDuABY/Yywins+4XjOzNunqaZDGx68PK9gXwLYZ12tm1iZdOg0SESOzLN/MrFyqvWed6QxGSf0kXSRpcrpdKKlflnWambVHtT98IOsnxVwDfAbsnW5zgWszrtPMrM3qo77krRKyzlmvFRF7Frw/S9KLGddpZtZm1T7dPOue9ReStm58k06S+SLjOs3M2qyBKHmrhKx71kcC1xfkqT8BDs64TjOzNqv2nnXWwfo14HxgLWA5YA7JQwleyrheM7M2qfbRIFkH678CnwLPA+9lXJeZWbt16XHWwGoRsVPGdZiZLbFqn26e9Q3Gf0j6esZ1mJktsYgoeauErHvWWwOHSHqbZG2QxiVSv5FxvWZmbdLVc9Y7Z1y+mVlZdOnRIBExPcvyzczKxY/1MjPLgS7dszYzy4tqHw3iYG1mhm8wmpnlgtMgZmY50NVnMJqZ5YJ71mZmOVDtOWtV+7eJgaSxETGu0u2w6uLfi64l67VBrDzGVroBVpX8e9GFOFibmeWAg7WZWQ44WOeD85LWHP9edCG+wWhmlgPuWZuZ5YCDtZlZDjhYVzlJ60l6UdILktbKoPx3JA0od7nWdpKOkfSapJvKXO4ISX8rZ5nW8TyDsfrtBvw1Is4o3ClJJPccqntdR2uLo4CdI+Ltxh2SukdEXQXbZFXCPesykbRG2iu6StIUSfdLWkrSxpKelvSSpDslLZ+e/4ik8yQ9K+lNScObKfO7wHHADyU9XFDHZcDzwOqSLpc0Oa3zrIJrF/eYJQ2T9Ej6un/athckXUnyXEyrMElXAGsCEyTNkTRO0v3ADenP/XFJz6fbluk1X+oxS7pE0iHp650kvS7pCWCPCnwkKzMH6/IaClwaERsAnwJ7AjcAp6QPCX4ZKOwhd4+IzUkC8hlNC4uIicAVwG8jYmS6e13ghojYJH1s2i8iYhjwDeA7klp7GPEZwBMRsQkwARjUzs9qZRQRRwAzgZHAb4FNgV0jYj/gQ2CHiPgmsA/w+2JlSeoNXAV8HxgOrJRh062DOFiX19sR8WL6+jlgLWC5iHg03Xc9sE3B+XcUnLtGiXVMj4inC97vLel54AVgA2D9Vq7fBhgPEBF3A5+UWK91rAkR8UX6ugdwlaSXgdto/We8Hsnv4luRjM0dn2E7rYM4Z11eCwte1wPLlXh+PenPQtK1wCbAzIj4bjPXzGt8IWkIcCKwWUR8Iuk6oHd6uI7/+zLuzZd5cH31m1fw+njgA2Ajkp/pgnR/4c8Yvvxz9s+4k3HPOltzgE8K8tEHAo8WOZ+IODQiNm4hUDfVl+T/1HMkfQ3YueDYOyR/SkOSjmn0GLA/gKSdgeVLqMcqqx/wfnoz+UCgJt0/HVhfUi9J/YDt0v2vA0MKRg+N6dDWWibcs87ewcAVkpYGpgGHlqvgiPinpBeAKWnZTxYcPgu4WtLPgWea7L85TZ08Cvy7XO2xzFwG/FnSD4CHSXvdEfGupFuBl4C3SFJhRMQCSWOBuyXNBp4ANqxIy61sPN3czCwHnAYxM8sBB2szsxxwsDYzywEHazOzHHCwNjPLAQdrQ1J9urLfK5JuS4cZtresxTRuUp8AAAN4SURBVOtVSBot6dQi5y4n6ah21HGmpBNbOHZQ+jmmSHq18TxJ10naq611mVULB2sD+CKdiLMhsAg4ovCgEm3+XYmICRFxbpFTliNZaa4s0kk+xwGj0vVZvkkyMcks9xysranHgbVbWOFvlKSn0pXfbpO0LLS8wpukQyRdkr7+Wrrq4D/TbUvgXGCttFd/QXreSZImpasUFq4i+AtJb0h6kGQxq+b8DDgxImZCMjkkIq5qepKk09M6XklXt1O6/5i0N/6SpD+l+76Ttq9xTfE+LbVT0jKS7k4/3yuS9lmCn4PZl3gGoy0mqTvJlPV7013rAodGxFHpcqunAdtHxDxJpwAnSDqfZIW3bYGpwC0tFP974NGI2F1SDbAscCqwYURsnNY/imTlws1Jlm6dIGkbkhl7+5KsmdKd5MvjuWbq2LCF/U1dEhFnp3XeCHwPuCttz5CIWCipcV2XE4GjI+LJ9MtpQZF2DiRZ02WXtOx+JbTFrCTuWRvAUpJeBCaTTD+/Ot1fuMLft0lWe3syPfdgYDClr/C2LXA5QETUR0Rz6YlR6fYCSUBejyQoDgfujIj5ETGXZGnXJTFS0jPpKnbbkqxWCMm07ZskHUCySBIkU/gvknQMyQqKdUXa+TKwvZJ1yoe38BnN2sU9a4M0Z124I80MFK78JuCBiBjT5LyNKd8KbwJ+HRFXNqnjuBLrmEKyeNXfW6wgWev5MmBYurbGmfzfanW7kCwhOxr4paQNIuJcSXcD3wWelrR9S+1My980PffXku5v7MGbLSn3rK1UTwNbSVobQNLSktah9BXeHgKOTK+tkdQX+AzoU3DOfcBhBbnwVSWtSLJS4O5KnrzTh2RR/eb8Gjhf0krp9b3SHnGhxsA8O61nr/TcbsDqEfEwcDLJzc9lJa0VES9HxHkkf3ms11I7Ja0CzI+I8cBvSG5wmpWFe9ZWkoiYpeSRUTdL6pXuPi0i3lRpK7wdC4yTdDjJ+t1HRsRTkp6U9ApwT0ScJOm/gKfSnv3nwAER8bykW4AXSZYFfbyFNk5UslTsg+lNwwCuaXLOp5KuIklZvANMSg/VAOPTPLNIns7zqaRzJI1M2/xq2s6FzbUTWBu4QFIDUEv65WRWDl51z8wsB5wGMTPLAQdrM7MccLA2M8sBB2szsxxwsDYzywEHazOzHHCwNjPLgf8P+hSJd6K5OnoAAAAASUVORK5CYII=\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plot_confusion_matrix(y_test, y_preds)"
]
},
{
"cell_type": "code",
"execution_count": 204,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" precision recall f1-score support\n",
"\n",
" non-fraud 1.00 1.00 1.00 28435\n",
" fraud 0.93 0.80 0.86 46\n",
"\n",
" accuracy 1.00 28481\n",
" macro avg 0.96 0.90 0.93 28481\n",
"weighted avg 1.00 1.00 1.00 28481\n",
"\n"
]
}
],
"source": [
"print(classification_report(\n",
" y_test, y_preds, target_names=['non-fraud', 'fraud']))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now that we have the sagemaker model, lets build the Amazon Fraud Detector one"
]
},
{
"cell_type": "code",
"execution_count": 91,
"metadata": {},
"outputs": [],
"source": [
"# -- This is all you need to fill out. Once complete simply interactively run each code cell. -- \n",
"# your_entity_name\n",
"ENTITY_TYPE = \"creditcardtrans{0}\".format(sufx) \n",
"ENTITY_DESC = \"creditcard transactions: {0}\".format(sufx) \n",
"# your_event_type\n",
"EVENT_TYPE = \"creditcard{0}\".format(sufx) \n",
"EVENT_DESC = \"creditcard card payment events: {0}\".format(sufx) \n",
"# your_model_name\n",
"MODEL_NAME = \"fraud_detector_model{0}\".format(sufx) \n",
"MODEL_DESC = \"model trained on: {0}\".format(sufx) \n",
"# your_detector_name\n",
"DETECTOR_NAME = \"fraud_detector_endpoint{0}\".format(sufx) \n",
"DETECTOR_DESC = \"detects synthetic fraud events created: {0}\".format(sufx) "
]
},
{
"cell_type": "code",
"execution_count": 92,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"--- summary stats ---\n",
" feature_name dtype count nunique null not_null null_pct nunique_pct feature_type feature_warning\n",
"0 time float64 256326 119735 0 256326 0.0 0.4671 NUMERIC NO WARNING\n",
"1 va float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"2 vb float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"3 vc float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"4 vd float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"5 ve float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"6 vf float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"7 vg float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"8 vh float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"9 vi float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"10 vj float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"11 vk float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"12 vl float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"13 vm float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"14 vn float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"15 vo float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"16 vp float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"17 vq float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"18 vr float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"19 vs float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"20 vt float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"21 vu float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"22 vv float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"23 vw float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"24 vx float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"25 vy float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"26 vz float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"27 vaa float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"28 vab float64 256326 248553 0 256326 0.0 0.9697 NUMERIC NO WARNING\n",
"29 amount float64 256326 31154 0 256326 0.0 0.1215 PRICE NO WARNING\n",
"30 EVENT_LABEL int64 256326 2 0 256326 0.0 0.0000 TARGET NO WARNING\n",
"31 EVENT_TIMESTAMP object 256326 119735 0 256326 0.0 0.4671 EVENT_TIMESTAMP NO WARNING\n",
"\n",
"\n",
"--- event variables ---\n",
"['time', 'va', 'vb', 'vc', 'vd', 've', 'vf', 'vg', 'vh', 'vi', 'vj', 'vk', 'vl', 'vm', 'vn', 'vo', 'vp', 'vq', 'vr', 'vs', 'vt', 'vu', 'vv', 'vw', 'vx', 'vy', 'vz', 'vaa', 'vab', 'amount']\n",
"\n",
"\n",
"--- event labels ---\n",
"[0, 1]\n",
"\n",
"\n",
"--- training data schema ---\n",
"{'modelVariables': ['time', 'va', 'vb', 'vc', 'vd', 've', 'vf', 'vg', 'vh', 'vi', 'vj', 'vk', 'vl', 'vm', 'vn', 'vo', 'vp', 'vq', 'vr', 'vs', 'vt', 'vu', 'vv', 'vw', 'vx', 'vy', 'vz', 'vaa', 'vab'], 'labelSchema': {'labelMapper': {'FRAUD': ['1'], 'LEGIT': ['0']}}}\n",
"\n",
"\n"
]
}
],
"source": [
"# Dataset profiling, just run this code block\n",
"def summary_stats(df):\n",
" \"\"\" Generate summary statistics for a panda's data frame \n",
" Args:\n",
" df (DataFrame): panda's dataframe to create summary statistics for.\n",
" Returns:\n",
" DataFrame of summary statistics, training data schema, event variables and event lables \n",
" \"\"\"\n",
" df = df.copy()\n",
" rowcnt = len(df)\n",
" df_s1 = df.agg(['count', 'nunique']).transpose().reset_index().rename(columns={\"index\":\"feature_name\"})\n",
" df_s1[\"null\"] = (rowcnt - df_s1[\"count\"]).astype('int64')\n",
" df_s1[\"not_null\"] = rowcnt - df_s1[\"null\"]\n",
" df_s1[\"null_pct\"] = df_s1[\"null\"] / rowcnt\n",
" df_s1[\"nunique_pct\"] = df_s1['nunique']/ rowcnt\n",
" dt = pd.DataFrame(df.dtypes).reset_index().rename(columns={\"index\":\"feature_name\", 0:\"dtype\"})\n",
" df_stats = pd.merge(dt, df_s1, on='feature_name', how='inner').round(4)\n",
" df_stats['nunique'] = df_stats['nunique'].astype('int64')\n",
" df_stats['count'] = df_stats['count'].astype('int64')\n",
" \n",
" # -- variable type mapper -- \n",
" df_stats['feature_type'] = \"UNKOWN\"\n",
" df_stats.loc[df_stats[\"dtype\"] == object, 'feature_type'] = \"CATEGORY\"\n",
" df_stats.loc[(df_stats[\"dtype\"] == \"int64\") | (df_stats[\"dtype\"] == \"float64\"), 'feature_type'] = \"NUMERIC\"\n",
" df_stats.loc[df_stats[\"feature_name\"].str.contains(\"ipaddress|ip_address|ipcli\"), 'feature_type'] = \"IP_ADDRESS\"\n",
" df_stats.loc[df_stats[\"feature_name\"].str.contains(\"email|email_address|emailaddr\"), 'feature_type'] = \"EMAIL_ADDRESS\"\n",
" df_stats.loc[df_stats[\"feature_name\"].str.contains(\"canal|channel\"), 'feature_type'] = \"USERAGENT\"\n",
" df_stats.loc[df_stats[\"feature_name\"].str.contains(\"monto|amount\"), 'feature_type'] = \"PRICE\"\n",
" df_stats.loc[df_stats[\"feature_name\"].str.contains(\"nomdes|name\"), 'feature_type'] = \"BILLING_NAME\"\n",
" df_stats.loc[df_stats[\"feature_name\"] == \"EVENT_LABEL\", 'feature_type'] = \"TARGET\"\n",
" df_stats.loc[df_stats[\"feature_name\"] == \"EVENT_TIMESTAMP\", 'feature_type'] = \"EVENT_TIMESTAMP\"\n",
" \n",
" # -- variable warnings -- \n",
" df_stats['feature_warning'] = \"NO WARNING\"\n",
" df_stats.loc[(df_stats[\"nunique\"] != 2) & (df_stats[\"feature_name\"] == \"EVENT_LABEL\"),'feature_warning' ] = \"LABEL WARNING, NON-BINARY EVENT LABEL\"\n",
" df_stats.loc[(df_stats[\"nunique_pct\"] > 0.97) & (df_stats['feature_type'] == \"CATEGORY\") ,'feature_warning' ] = \"EXCLUDE, GT 97% UNIQUE\"\n",
" df_stats.loc[(df_stats[\"null_pct\"] > 0.2) & (df_stats[\"null_pct\"] <= 0.5), 'feature_warning' ] = \"NULL WARNING, GT 20% MISSING\"\n",
" df_stats.loc[df_stats[\"null_pct\"] > 0.5,'feature_warning' ] = \"EXCLUDE, GT 50% MISSING\"\n",
" df_stats.loc[((df_stats['dtype'] == \"int64\" ) | (df_stats['dtype'] == \"float64\" ) ) & (df_stats['nunique'] < 0.2), 'feature_warning' ] = \"LIKELY CATEGORICAL, NUMERIC w. LOW CARDINALITY\"\n",
" \n",
" # -- target check -- \n",
" exclude_fields = df_stats.loc[(df_stats['feature_warning'] != 'NO WARNING')]['feature_name'].to_list()\n",
" event_variables = df_stats.loc[(~df_stats['feature_name'].isin(['EVENT_LABEL', 'EVENT_TIMESTAMP']))]['feature_name'].to_list()\n",
" event_labels = df[\"EVENT_LABEL\"].unique().tolist()\n",
" \n",
" trainingDataSchema = {\n",
" 'modelVariables' : df_stats.loc[(df_stats['feature_type'].isin(['IP_ADDRESS', 'EMAIL_ADDRESS', 'CATEGORY', 'NUMERIC' ]))]['feature_name'].to_list(),\n",
" 'labelSchema' : {\n",
" 'labelMapper' : {\n",
" 'FRAUD' : [str(df[\"EVENT_LABEL\"].value_counts().idxmin())],\n",
" 'LEGIT' : [str(df[\"EVENT_LABEL\"].value_counts().idxmax())]\n",
" }\n",
" }\n",
" }\n",
" \n",
" \n",
" model_variables = df_stats.loc[(df_stats['feature_type'].isin(['IP_ADDRESS', 'EMAIL_ADDRESS', 'CATEGORY', 'NUMERIC' ]))]['feature_name'].to_list()\n",
" \n",
" \n",
" # -- label schema -- \n",
" label_map = {\n",
" 'FRAUD' : [df[\"EVENT_LABEL\"].value_counts().idxmin()],\n",
" 'LEGIT' : [df[\"EVENT_LABEL\"].value_counts().idxmax()]\n",
" }\n",
" \n",
" \n",
" print(\"--- summary stats ---\")\n",
" print(df_stats)\n",
" print(\"\\n\")\n",
" print(\"--- event variables ---\")\n",
" print(event_variables)\n",
" print(\"\\n\")\n",
" print(\"--- event labels ---\")\n",
" print(event_labels)\n",
" print(\"\\n\")\n",
" print(\"--- training data schema ---\")\n",
" print(trainingDataSchema)\n",
" print(\"\\n\")\n",
" \n",
" return df_stats, trainingDataSchema, event_variables, event_labels\n",
"\n",
"# -- connect to S3, snag file, and convert to a panda's dataframe --\n",
"#s3 = boto3.resource('s3')\n",
"#obj = s3.Object(S3_BUCKET, S3_FILE)\n",
"#body = obj.get()['Body']\n",
"#df = pd.read_csv(body)\n",
"\n",
"# -- call profiling function -- \n",
"df_stats, trainingDataSchema, eventVariables, eventLabels = summary_stats(data)"
]
},
{
"cell_type": "code",
"execution_count": 93,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Creating variable: time\n",
"Creating variable: fraud_detector_model20200922_insightscore\n",
"\n",
" --- model variable dict --\n",
"[{'name': 'amount'}, {'name': 'time'}, {'name': 'va'}, {'name': 'vb'}, {'name': 'vc'}, {'name': 'vd'}, {'name': 've'}, {'name': 'vf'}, {'name': 'vg'}, {'name': 'vh'}, {'name': 'vi'}, {'name': 'vj'}, {'name': 'vk'}, {'name': 'vl'}, {'name': 'vm'}, {'name': 'vn'}, {'name': 'vo'}, {'name': 'vp'}, {'name': 'vq'}, {'name': 'vr'}, {'name': 'vs'}, {'name': 'vt'}, {'name': 'vu'}, {'name': 'vv'}, {'name': 'vw'}, {'name': 'vx'}, {'name': 'vy'}, {'name': 'vz'}, {'name': 'vaa'}, {'name': 'vab'}]\n",
"\n",
" --- model label schema dict --\n",
"{'labelKey': 'EVENT_LABEL', 'labelMapper': {'FRAUD': ['1'], 'LEGIT': ['0']}}\n"
]
}
],
"source": [
"# Variable creation just run this code block\n",
"def create_label(df, FRAUD_LABEL):\n",
" \"\"\"\n",
" Returns a dictionary for the model labelSchema, by identifying the rare event as fraud / and common as not-fraud \n",
" \n",
" Arguments:\n",
" df -- input dataframe \n",
" FRAUD_LABEL -- the name of the field that contains fraud label \n",
" \n",
" Returns:\n",
" labelSchema -- a dictionary containing labelKey & labelMapper \n",
" \"\"\"\n",
" label_summary = df[FRAUD_LABEL].value_counts()\n",
" labelSchema = {'labelKey': FRAUD_LABEL,\n",
" \"labelMapper\" : { \"FRAUD\": [str(label_summary.idxmin())], \n",
" \"LEGIT\": [str(label_summary.idxmax())]}\n",
" }\n",
" afd_resource.put_label(\n",
" name = str(label_summary.idxmin()),\n",
" description = 'FRAUD')\n",
" \n",
" afd_resource.put_label(\n",
" name = str(label_summary.idxmax()),\n",
" description = 'LEGIT')\n",
" \n",
" return labelSchema\n",
" \n",
"# -- function to create all your variables --- \n",
"def create_variables(df_stats, MODEL_NAME):\n",
" \"\"\"\n",
" Returns a variable list of model input variables, checks to see if variable exists,\n",
" and, if not, then it adds the variable to Fraud Detector \n",
" \n",
" Arguments: \n",
" enrichment_features -- dictionary of optional features, mapped to specific variable types enriched (CARD_BIN, USERAGENT)\n",
" numeric_features -- optional list of numeric field names \n",
" categorical_features -- optional list of categorical features \n",
" \n",
" Returns:\n",
" variable_list -- a list of variable dictionaries \n",
" \n",
" \"\"\"\n",
" enrichment_features = df_stats.loc[(df_stats['feature_type'].isin(['IP_ADDRESS', 'EMAIL_ADDRESS', 'USERAGENT', 'BILLING_NAME', 'PRICE']))]['feature_name'].to_dict()\n",
" enrichment_type = df_stats.loc[(df_stats['feature_type'].isin(['IP_ADDRESS', 'EMAIL_ADDRESS', 'USERAGENT', 'BILLING_NAME', 'PRICE']))]['feature_type'].to_dict()\n",
" numeric_features = df_stats.loc[(df_stats['feature_type'].isin(['NUMERIC']))]['feature_name'].to_dict()\n",
" categorical_features = df_stats.loc[(df_stats['feature_type'].isin(['CATEGORY']))]['feature_name'].to_dict()\n",
" \n",
" variable_list = []\n",
" # -- first do the enrichment features\n",
" for feature in enrichment_features.keys(): \n",
" variable_list.append( {'name' : enrichment_features[feature]+\"\"})\n",
" try:\n",
" varname = enrichment_features[feature]+\"\"\n",
" afd_resource.get_variables(name=varname)\n",
" except:\n",
" print(\"Creating variable: {0}\".format(enrichment_features[feature]))\n",
" if enrichment_type[feature] == \"PRICE\":\n",
" resp = afd_resource.create_variable(\n",
" name = varname,\n",
" dataType = 'FLOAT',\n",
" dataSource ='EVENT',\n",
" defaultValue = '0', \n",
" description = enrichment_features[feature],\n",
" variableType = enrichment_type[feature] )\n",
" else:\n",
" resp = afd_resource.create_variable(\n",
" name = varname,\n",
" dataType = 'STRING',\n",
" dataSource ='EVENT',\n",
" defaultValue = '', \n",
" description = enrichment_features[feature],\n",
" variableType = enrichment_type[feature] )\n",
" \n",
" \n",
" # -- check and update the numeric features \n",
" for feature in numeric_features: \n",
" variable_list.append( {'name' : numeric_features[feature]+\"\"})\n",
" try:\n",
" varname = numeric_features[feature]+\"\"\n",
" afd_resource.get_variables(name=varname)\n",
" except:\n",
" print(\"Creating variable: {0}\".format(numeric_features[feature]))\n",
" resp = afd_resource.create_variable(\n",
" name = varname,\n",
" dataType = 'FLOAT',\n",
" dataSource ='EVENT',\n",
" defaultValue = '0.0', \n",
" description = numeric_features[feature],\n",
" variableType = 'NUMERIC' )\n",
" \n",
" # -- check and update the categorical features \n",
" for feature in categorical_features: \n",
" variable_list.append( {'name' : categorical_features[feature]+\"\"})\n",
" try:\n",
" varname = categorical_features[feature]+\"\"\n",
" afd_resource.get_variables(name=varname)\n",
" except:\n",
" print(\"Creating variable: {0}\".format(categorical_features[feature]))\n",
" resp = afd_resource.create_variable(\n",
" name = varname,\n",
" dataType = 'STRING',\n",
" dataSource ='EVENT',\n",
" defaultValue = '', \n",
" description = categorical_features[feature],\n",
" variableType = 'CATEGORICAL' )\n",
" \n",
" # -- create a model score feature \n",
" model_feature = \"{0}_insightscore\".format(MODEL_NAME) \n",
" # variable_list.append( {'name' : model_feature})\n",
" try:\n",
" afd_resource.get_variables(name=model_feature)\n",
" except:\n",
" print(\"Creating variable: {0}\".format(model_feature))\n",
" resp = afd_resource.create_variable(\n",
" name = model_feature,\n",
" dataType = 'FLOAT',\n",
" dataSource ='MODEL_SCORE',\n",
" defaultValue = '0.0', \n",
" description = model_feature,\n",
" variableType = 'NUMERIC' )\n",
" \n",
" return variable_list\n",
"\n",
"\n",
"model_variables = create_variables(df_stats, MODEL_NAME)\n",
"print(\"\\n --- model variable dict --\")\n",
"print(model_variables)\n",
"\n",
"\n",
"model_label = create_label(data, \"EVENT_LABEL\")\n",
"print(\"\\n --- model label schema dict --\")\n",
"print(model_label)"
]
},
{
"cell_type": "code",
"execution_count": 94,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['0', '1']\n"
]
}
],
"source": [
"#Amazon Fraud Detector expect the labels to be strings.\n",
"eventLabels = list(map(str, eventLabels))\n",
"print(eventLabels)"
]
},
{
"cell_type": "code",
"execution_count": 95,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"-- create entity --\n",
"{'ResponseMetadata': {'RequestId': '88e9d45f-2c8b-4a90-b58c-8a6dcea3bc40', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 21:48:52 GMT', 'x-amzn-requestid': '88e9d45f-2c8b-4a90-b58c-8a6dcea3bc40', 'content-length': '2', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n",
"-- create event type --\n",
"{'ResponseMetadata': {'RequestId': 'd48b2c1c-34a2-4c97-bc7d-52c0e192f729', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 21:48:53 GMT', 'x-amzn-requestid': 'd48b2c1c-34a2-4c97-bc7d-52c0e192f729', 'content-length': '2', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# Creating entity and event types just run this code block ---\n",
"response = afd_resource.put_entity_type(\n",
" name = ENTITY_TYPE,\n",
" description = ENTITY_DESC\n",
")\n",
"print(\"-- create entity --\")\n",
"print(response)\n",
"\n",
"response = afd_resource.put_event_type (\n",
" name = EVENT_TYPE,\n",
" eventVariables = eventVariables,\n",
" labels = eventLabels,\n",
" entityTypes = [ENTITY_TYPE])\n",
"print(\"-- create event type --\")\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 96,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"-- initalize model --\n",
"{'ResponseMetadata': {'RequestId': '2c15f806-c600-4217-b400-8332dda66a47', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 21:48:56 GMT', 'x-amzn-requestid': '2c15f806-c600-4217-b400-8332dda66a47', 'content-length': '2', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# -- create our model --\n",
"response = afd_resource.create_model(\n",
" description = MODEL_DESC,\n",
" eventTypeName = EVENT_TYPE,\n",
" modelId = MODEL_NAME,\n",
" modelType = 'ONLINE_FRAUD_INSIGHTS')\n",
"print(\"-- initalize model --\")\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 97,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"-- model training --\n",
"{'modelId': 'fraud_detector_model20200922', 'modelType': 'ONLINE_FRAUD_INSIGHTS', 'modelVersionNumber': '1.0', 'status': 'TRAINING_IN_PROGRESS', 'ResponseMetadata': {'RequestId': 'e88dc617-42fe-4971-9509-ee6d1fbdf6f8', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 21:49:02 GMT', 'x-amzn-requestid': 'e88dc617-42fe-4971-9509-ee6d1fbdf6f8', 'content-length': '137', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# -- initializes the model, it's now ready to train -- \n",
"S3_FILE = \"dataset-training.csv\"\n",
"S3_FILE_LOC = \"s3://{0}/{1}\".format(S3_BUCKET,S3_FILE)\n",
"\n",
"response = afd_resource.create_model_version(\n",
" modelId = MODEL_NAME,\n",
" modelType = 'ONLINE_FRAUD_INSIGHTS',\n",
" trainingDataSource = 'EXTERNAL_EVENTS',\n",
" trainingDataSchema = trainingDataSchema,\n",
" externalEventsDetail = {\n",
" 'dataLocation' : S3_FILE_LOC,\n",
" 'dataAccessRoleArn': ARN_ROLE\n",
" }\n",
")\n",
"print(\"-- model training --\")\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 98,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model status : TRAINING_COMPLETE\n",
"\n",
" --- model training complete --\n",
"Elapsed time : 3134.791792869568 seconds \n",
"\n",
"{'modelId': 'fraud_detector_model20200922', 'modelType': 'ONLINE_FRAUD_INSIGHTS', 'modelVersionNumber': '1.0', 'trainingDataSource': 'EXTERNAL_EVENTS', 'trainingDataSchema': {'modelVariables': ['time', 'va', 'vb', 'vc', 'vd', 've', 'vf', 'vg', 'vh', 'vi', 'vj', 'vk', 'vl', 'vm', 'vn', 'vo', 'vp', 'vq', 'vr', 'vs', 'vt', 'vu', 'vv', 'vw', 'vx', 'vy', 'vz', 'vaa', 'vab'], 'labelSchema': {'labelMapper': {'FRAUD': ['1'], 'LEGIT': ['0']}}}, 'externalEventsDetail': {'dataLocation': 's3://afd-poc-trainingbucket-1i37svk9elcoe/dataset-training.csv', 'dataAccessRoleArn': 'arn:aws:iam::387461613214:role/afd-poc-NotebookInstanceExecutionRole-1FNQ41S8H2G68'}, 'status': 'TRAINING_COMPLETE', 'arn': 'arn:aws:frauddetector:us-west-2:387461613214:model-version/ONLINE_FRAUD_INSIGHTS/fraud_detector_model20200922/1.0', 'ResponseMetadata': {'RequestId': '7386e6d4-a8b0-4033-b965-899cd5814a26', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 22:41:31 GMT', 'x-amzn-requestid': '7386e6d4-a8b0-4033-b965-899cd5814a26', 'content-length': '756', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# -- model training takes time, we'll loop until it's complete -- \n",
"print(\"-- wait for model training to complete --\")\n",
"stime = time.time()\n",
"while True:\n",
" clear_output(wait=True)\n",
" response = afd_resource.get_model_version(modelId=MODEL_NAME, modelType = \"ONLINE_FRAUD_INSIGHTS\", modelVersionNumber = '1.0')\n",
" if response['status'] == 'TRAINING_IN_PROGRESS':\n",
" print(f\"current progress: {(time.time() - stime)/60:{3}.{3}} minutes\")\n",
" time.sleep(60) # -- sleep for 60 seconds \n",
" if response['status'] != 'TRAINING_IN_PROGRESS':\n",
" print(\"Model status : \" + response['status'])\n",
" break\n",
" \n",
"etime = time.time()\n",
"\n",
"# -- summarize -- \n",
"print(\"\\n --- model training complete --\")\n",
"print(\"Elapsed time : %s\" % (etime - stime) + \" seconds \\n\" )\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 99,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model status : ACTIVE\n",
"Elapsed time : 602.6856355667114 seconds \n",
"\n",
"{'modelId': 'fraud_detector_model20200922', 'modelType': 'ONLINE_FRAUD_INSIGHTS', 'modelVersionNumber': '1.0', 'trainingDataSource': 'EXTERNAL_EVENTS', 'trainingDataSchema': {'modelVariables': ['time', 'va', 'vb', 'vc', 'vd', 've', 'vf', 'vg', 'vh', 'vi', 'vj', 'vk', 'vl', 'vm', 'vn', 'vo', 'vp', 'vq', 'vr', 'vs', 'vt', 'vu', 'vv', 'vw', 'vx', 'vy', 'vz', 'vaa', 'vab'], 'labelSchema': {'labelMapper': {'FRAUD': ['1'], 'LEGIT': ['0']}}}, 'externalEventsDetail': {'dataLocation': 's3://afd-poc-trainingbucket-1i37svk9elcoe/dataset-training.csv', 'dataAccessRoleArn': 'arn:aws:iam::387461613214:role/afd-poc-NotebookInstanceExecutionRole-1FNQ41S8H2G68'}, 'status': 'ACTIVE', 'arn': 'arn:aws:frauddetector:us-west-2:387461613214:model-version/ONLINE_FRAUD_INSIGHTS/fraud_detector_model20200922/1.0', 'ResponseMetadata': {'RequestId': 'bca2079c-03d0-42f1-9863-7cd0c71fc4dc', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 22:54:49 GMT', 'x-amzn-requestid': 'bca2079c-03d0-42f1-9863-7cd0c71fc4dc', 'content-length': '745', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# activating the model\n",
"response = afd_resource.update_model_version_status (\n",
" modelId = MODEL_NAME,\n",
" modelType = 'ONLINE_FRAUD_INSIGHTS',\n",
" modelVersionNumber = '1.0',\n",
" status = 'ACTIVE'\n",
")\n",
"print(\"-- activating model --\")\n",
"print(response)\n",
"\n",
"#-- wait until model is active \n",
"print(\"--- waiting until model status is active \")\n",
"stime = time.time()\n",
"while True:\n",
" clear_output(wait=True)\n",
" response = afd_resource.get_model_version(modelId=MODEL_NAME, modelType = \"ONLINE_FRAUD_INSIGHTS\", modelVersionNumber = '1.0')\n",
" if response['status'] != 'ACTIVE':\n",
" print(f\"current progress: {(time.time() - stime)/60:{3}.{3}} minutes\")\n",
" time.sleep(60) # sleep for 1 minute \n",
" if response['status'] == 'ACTIVE':\n",
" print(\"Model status : \" + response['status'])\n",
" break\n",
" \n",
"etime = time.time()\n",
"print(\"Elapsed time : %s\" % (etime - stime) + \" seconds \\n\" )\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 185,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# -- model performance summary -- \n",
"auc = afd_resource.describe_model_versions(\n",
" modelId= MODEL_NAME,\n",
" modelVersionNumber='1.0',\n",
" modelType='ONLINE_FRAUD_INSIGHTS',\n",
" maxResults=10\n",
")['modelVersionDetails'][0]['trainingResult']['trainingMetrics']['auc']\n",
"\n",
"\n",
"df_model = pd.DataFrame(afd_resource.describe_model_versions(\n",
" modelId= MODEL_NAME,\n",
" modelVersionNumber='1.0',\n",
" modelType='ONLINE_FRAUD_INSIGHTS',\n",
" maxResults=10\n",
")['modelVersionDetails'][0]['trainingResult']['trainingMetrics']['metricDataPoints'])\n",
"\n",
"\n",
"plt.figure(figsize=(6,6))\n",
"plt.plot(df_model[\"fpr\"], df_model[\"tpr\"], color='darkorange',\n",
" lw=2, label='ROC curve (area = %0.3f)' % auc)\n",
"plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')\n",
"plt.xlabel('False Positive Rate')\n",
"plt.ylabel('True Positive Rate')\n",
"plt.title( MODEL_NAME + ' ROC Chart')\n",
"plt.legend(loc=\"lower right\",fontsize=12)\n",
"plt.axvline(x = 0.02 ,linewidth=2, color='r')\n",
"plt.axhline(y = 0.73 ,linewidth=2, color='r')\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 101,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'ResponseMetadata': {'RequestId': '4c1eb5b8-0bb0-4ac0-abf0-49d55c3861c6', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 22:55:06 GMT', 'x-amzn-requestid': '4c1eb5b8-0bb0-4ac0-abf0-49d55c3861c6', 'content-length': '2', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# -- initialize your detector -- \n",
"response = afd_resource.put_detector(detectorId = DETECTOR_NAME, \n",
" description = DETECTOR_DESC, \n",
" eventTypeName = EVENT_TYPE )\n",
"\n",
"print(response)"
]
},
{
"cell_type": "code",
"execution_count": 102,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" --- score thresholds 1% to 6% --- \n",
" fpr tpr threshold rule outcome\n",
"0 0.01 0.85 970.0 $fraud_detector_model20200922_insightscore > 9... fraud\n",
"1 0.02 0.87 925.0 $fraud_detector_model20200922_insightscore > 9... fraud\n",
"2 0.03 0.87 880.0 $fraud_detector_model20200922_insightscore > 8... fraud\n",
"3 0.04 0.90 835.0 $fraud_detector_model20200922_insightscore > 8... investigate\n",
"4 0.05 0.91 790.0 $fraud_detector_model20200922_insightscore > 7... investigate\n",
"5 0.06 0.91 755.0 $fraud_detector_model20200922_insightscore <= ... approve\n"
]
}
],
"source": [
"# -- make rules -- \n",
"model_stat = df_model.round(decimals=2) \n",
"\n",
"m = model_stat.loc[model_stat.groupby([\"fpr\"])[\"threshold\"].idxmax()] \n",
"\n",
"def make_rule(x):\n",
" rule = \"\"\n",
" if x['fpr'] <= 0.05: \n",
" rule = \"${0}_insightscore > {1}\".format(MODEL_NAME,x['threshold'])\n",
" if x['fpr'] == 0.06:\n",
" rule = \"${0}_insightscore <= {1}\".format(MODEL_NAME,x['threshold_prev'])\n",
" return rule\n",
" \n",
"m[\"threshold_prev\"] = m['threshold'].shift(1)\n",
"m['rule'] = m.apply(lambda x: make_rule(x), axis=1)\n",
"\n",
"m['outcome'] = \"approve\"\n",
"m.loc[m['fpr'] <= 0.03, \"outcome\"] = \"fraud\"\n",
"m.loc[(m['fpr'] > 0.03) & (m['fpr'] <= 0.05), \"outcome\"] = \"investigate\"\n",
"\n",
"print (\" --- score thresholds 1% to 6% --- \")\n",
"print(m[[\"fpr\", \"tpr\", \"threshold\", \"rule\", \"outcome\"]].loc[(m['fpr'] > 0.0 ) & (m['fpr'] <= 0.06)].reset_index(drop=True))"
]
},
{
"cell_type": "code",
"execution_count": 103,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"creating outcome variable: fraud \n",
"creating outcome variable: investigate \n",
"creating outcome variable: approve \n"
]
}
],
"source": [
"# -- create outcomes -- \n",
"def create_outcomes(outcomes):\n",
" \"\"\" create Fraud Detector Outcomes \n",
" \n",
" \"\"\" \n",
" for outcome in outcomes:\n",
" print(\"creating outcome variable: {0} \".format(outcome))\n",
" response = afd_resource.put_outcome(\n",
" name=outcome,\n",
" description=outcome)\n",
"\n",
"# -- get distinct outcomes \n",
"outcomes = m[\"outcome\"].unique().tolist()\n",
"\n",
"create_outcomes(outcomes)"
]
},
{
"cell_type": "code",
"execution_count": 104,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"creating rule: rule0_fraud_detector_model20200922: IF $fraud_detector_model20200922_insightscore > 970.0 THEN fraud\n",
"creating rule: rule1_fraud_detector_model20200922: IF $fraud_detector_model20200922_insightscore > 925.0 THEN fraud\n",
"creating rule: rule2_fraud_detector_model20200922: IF $fraud_detector_model20200922_insightscore > 880.0 THEN fraud\n",
"creating rule: rule3_fraud_detector_model20200922: IF $fraud_detector_model20200922_insightscore > 835.0 THEN investigate\n",
"creating rule: rule4_fraud_detector_model20200922: IF $fraud_detector_model20200922_insightscore > 790.0 THEN investigate\n",
"creating rule: rule5_fraud_detector_model20200922: IF $fraud_detector_model20200922_insightscore <= 790.0 THEN approve\n"
]
},
{
"data": {
"text/plain": [
"[{'ruleId': 'rule0_fraud_detector_model20200922',\n",
" 'ruleVersion': '1',\n",
" 'detectorId': 'fraud_detector_endpoint20200922'},\n",
" {'ruleId': 'rule1_fraud_detector_model20200922',\n",
" 'ruleVersion': '1',\n",
" 'detectorId': 'fraud_detector_endpoint20200922'},\n",
" {'ruleId': 'rule2_fraud_detector_model20200922',\n",
" 'ruleVersion': '1',\n",
" 'detectorId': 'fraud_detector_endpoint20200922'},\n",
" {'ruleId': 'rule3_fraud_detector_model20200922',\n",
" 'ruleVersion': '1',\n",
" 'detectorId': 'fraud_detector_endpoint20200922'},\n",
" {'ruleId': 'rule4_fraud_detector_model20200922',\n",
" 'ruleVersion': '1',\n",
" 'detectorId': 'fraud_detector_endpoint20200922'},\n",
" {'ruleId': 'rule5_fraud_detector_model20200922',\n",
" 'ruleVersion': '1',\n",
" 'detectorId': 'fraud_detector_endpoint20200922'}]"
]
},
"execution_count": 104,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# creating te rules with the desired outcomes associated with our detector\n",
"rule_set = m[(m[\"fpr\"] > 0.0) & (m[\"fpr\"] <= 0.06)][[\"outcome\", \"rule\"]].to_dict('records')\n",
"rule_list = []\n",
"for i, rule in enumerate(rule_set):\n",
" ruleId = \"rule{0}_{1}\".format(i, MODEL_NAME)\n",
" rule_list.append({\"ruleId\": ruleId, \n",
" \"ruleVersion\" : '1',\n",
" \"detectorId\" : DETECTOR_NAME\n",
" \n",
" })\n",
" print(\"creating rule: {0}: IF {1} THEN {2}\".format(ruleId, rule[\"rule\"], rule['outcome']))\n",
" try:\n",
" response = afd_resource.create_rule(\n",
" ruleId = ruleId,\n",
" detectorId = DETECTOR_NAME,\n",
" expression = rule['rule'],\n",
" language = 'DETECTORPL',\n",
" outcomes = [rule['outcome']]\n",
" )\n",
" except:\n",
" print(\"this rule already exists in this detector\")\n",
"rule_list"
]
},
{
"cell_type": "code",
"execution_count": 105,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
" -- detector created -- \n",
"{'rule': {'detectorId': 'fraud_detector_endpoint20200922', 'ruleId': 'rule5_fraud_detector_model20200922', 'ruleVersion': '1'}, 'ResponseMetadata': {'RequestId': 'ceb37fe4-61df-4797-b198-341bf37b8a7d', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 22:55:26 GMT', 'x-amzn-requestid': 'ceb37fe4-61df-4797-b198-341bf37b8a7d', 'content-length': '121', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# creating the detector\n",
"afd_resource.create_detector_version(\n",
" detectorId = DETECTOR_NAME,\n",
" rules = rule_list,\n",
" modelVersions = [{\"modelId\":MODEL_NAME, \n",
" \"modelType\" : \"ONLINE_FRAUD_INSIGHTS\",\n",
" \"modelVersionNumber\" : \"1.0\"}],\n",
" ruleExecutionMode = 'FIRST_MATCHED'\n",
" )\n",
"\n",
"print(\"\\n -- detector created -- \")\n",
"print(response) "
]
},
{
"cell_type": "code",
"execution_count": 106,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
" -- detector activated -- \n",
"{'ResponseMetadata': {'RequestId': '12c43795-ac18-42d6-8d07-22f49ca7014d', 'HTTPStatusCode': 200, 'HTTPHeaders': {'content-type': 'application/x-amz-json-1.1', 'date': 'Tue, 22 Sep 2020 22:55:37 GMT', 'x-amzn-requestid': '12c43795-ac18-42d6-8d07-22f49ca7014d', 'content-length': '2', 'connection': 'keep-alive'}, 'RetryAttempts': 0}}\n"
]
}
],
"source": [
"# activating the detector\n",
"response = afd_resource.update_detector_version_status(\n",
" detectorId= DETECTOR_NAME,\n",
" detectorVersionId='1',\n",
" status='ACTIVE'\n",
")\n",
"print(\"\\n -- detector activated -- \")\n",
"print(response)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Testing our model endpoint"
]
},
{
"cell_type": "code",
"execution_count": 107,
"metadata": {},
"outputs": [],
"source": [
"S3_FILE = \"dataset-test.csv\"\n",
"S3_FILE_LOC = \"s3://{0}/{1}\".format(S3_BUCKET,S3_FILE)\n",
"\n",
"s3_resource.Bucket(S3_BUCKET).download_file(S3_FILE, 'dataset-test.csv')"
]
},
{
"cell_type": "code",
"execution_count": 108,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 41505.0 \n",
" -16.526507 \n",
" 8.584972 \n",
" -18.649853 \n",
" 9.505594 \n",
" -13.793819 \n",
" -2.832404 \n",
" -16.701694 \n",
" 7.517344 \n",
" -8.507059 \n",
" -14.110184 \n",
" 5.299236 \n",
" -10.834006 \n",
" 1.671120 \n",
" -9.373859 \n",
" 0.360806 \n",
" -9.899247 \n",
" -19.236292 \n",
" -8.398552 \n",
" 3.101735 \n",
" -1.514923 \n",
" 1.190739 \n",
" -1.127670 \n",
" -2.358579 \n",
" 0.673461 \n",
" -1.413700 \n",
" -0.462762 \n",
" -2.018575 \n",
" -1.042804 \n",
" 364.19 \n",
" \n",
" \n",
" 1 \n",
" 44261.0 \n",
" 0.339812 \n",
" -2.743745 \n",
" -0.134070 \n",
" -1.385729 \n",
" -1.451413 \n",
" 1.015887 \n",
" -0.524379 \n",
" 0.224060 \n",
" 0.899746 \n",
" -0.565012 \n",
" -0.087670 \n",
" 0.979427 \n",
" 0.076883 \n",
" -0.217884 \n",
" -0.136830 \n",
" -2.142892 \n",
" 0.126956 \n",
" 1.752662 \n",
" 0.432546 \n",
" 0.506044 \n",
" -0.213436 \n",
" -0.942525 \n",
" -0.526819 \n",
" -1.156992 \n",
" 0.311211 \n",
" -0.746647 \n",
" 0.040996 \n",
" 0.102038 \n",
" 520.12 \n",
" \n",
" \n",
" 2 \n",
" 35484.0 \n",
" 1.399590 \n",
" -0.590701 \n",
" 0.168619 \n",
" -1.029950 \n",
" -0.539806 \n",
" 0.040444 \n",
" -0.712567 \n",
" 0.002299 \n",
" -0.971747 \n",
" 0.756801 \n",
" 0.543827 \n",
" 0.112453 \n",
" 1.075384 \n",
" -0.245772 \n",
" 0.180483 \n",
" 1.769860 \n",
" -0.533172 \n",
" -0.533300 \n",
" 1.192245 \n",
" 0.212877 \n",
" 0.102398 \n",
" 0.168269 \n",
" -0.166639 \n",
" -0.810250 \n",
" 0.505083 \n",
" -0.232340 \n",
" 0.011409 \n",
" 0.004634 \n",
" 31.00 \n",
" \n",
" \n",
" 3 \n",
" 167123.0 \n",
" -0.432071 \n",
" 1.647895 \n",
" -1.669361 \n",
" -0.349504 \n",
" 0.785785 \n",
" -0.630647 \n",
" 0.276990 \n",
" 0.586025 \n",
" -0.484715 \n",
" -1.376648 \n",
" -1.328335 \n",
" 0.223621 \n",
" 1.132627 \n",
" -0.550875 \n",
" 0.616568 \n",
" 0.497974 \n",
" 0.502195 \n",
" 0.981343 \n",
" 0.101264 \n",
" -0.244633 \n",
" 0.358932 \n",
" 0.873663 \n",
" -0.178642 \n",
" -0.017171 \n",
" -0.207392 \n",
" -0.157756 \n",
" -0.237386 \n",
" 0.001934 \n",
" 1.50 \n",
" \n",
" \n",
" 4 \n",
" 168473.0 \n",
" 2.014160 \n",
" -0.137394 \n",
" -1.015839 \n",
" 0.327269 \n",
" -0.182179 \n",
" -0.956571 \n",
" 0.043241 \n",
" -0.160746 \n",
" 0.363241 \n",
" 0.259452 \n",
" 0.942162 \n",
" 0.850038 \n",
" -0.616166 \n",
" 0.592634 \n",
" -0.603845 \n",
" 0.091077 \n",
" -0.471867 \n",
" -0.333816 \n",
" 0.404711 \n",
" -0.255293 \n",
" -0.238644 \n",
" -0.616400 \n",
" 0.347045 \n",
" 0.061561 \n",
" -0.360196 \n",
" 0.174730 \n",
" -0.078043 \n",
" -0.070571 \n",
" 0.89 \n",
" \n",
" \n",
" 5 \n",
" 67878.0 \n",
" -0.641330 \n",
" -0.057304 \n",
" 1.489998 \n",
" -1.688131 \n",
" -1.151043 \n",
" 0.259996 \n",
" -1.391069 \n",
" -2.334075 \n",
" 1.168644 \n",
" -2.084080 \n",
" 0.480381 \n",
" 0.473738 \n",
" -2.192276 \n",
" 0.773942 \n",
" 0.294484 \n",
" 0.406074 \n",
" -0.541855 \n",
" 1.031450 \n",
" 0.017076 \n",
" 0.618411 \n",
" -1.231634 \n",
" 0.257164 \n",
" -0.371953 \n",
" -0.038566 \n",
" 1.397514 \n",
" -0.665947 \n",
" 0.031003 \n",
" 0.180357 \n",
" 100.00 \n",
" \n",
" \n",
" 6 \n",
" 159763.0 \n",
" 2.023952 \n",
" -0.120140 \n",
" -1.086918 \n",
" 0.423019 \n",
" -0.142901 \n",
" -1.127752 \n",
" 0.178493 \n",
" -0.303234 \n",
" 0.564509 \n",
" 0.062831 \n",
" -0.720047 \n",
" 0.366835 \n",
" -0.110857 \n",
" 0.319094 \n",
" 0.108359 \n",
" -0.153633 \n",
" -0.221312 \n",
" -0.934141 \n",
" 0.070553 \n",
" -0.210864 \n",
" -0.276175 \n",
" -0.697708 \n",
" 0.335631 \n",
" -0.017196 \n",
" -0.324904 \n",
" 0.200023 \n",
" -0.071566 \n",
" -0.058224 \n",
" 16.99 \n",
" \n",
" \n",
" 7 \n",
" 139631.0 \n",
" -0.688944 \n",
" 1.292153 \n",
" -0.564281 \n",
" -1.457526 \n",
" 2.258333 \n",
" -0.323270 \n",
" 1.678984 \n",
" -0.104128 \n",
" -1.285351 \n",
" -1.303435 \n",
" 0.282728 \n",
" -0.402525 \n",
" -0.548687 \n",
" -0.504283 \n",
" -0.685339 \n",
" 0.714828 \n",
" -0.092674 \n",
" 0.798953 \n",
" -0.150085 \n",
" -0.037150 \n",
" -0.006880 \n",
" -0.171568 \n",
" -0.720019 \n",
" -0.419435 \n",
" 1.211991 \n",
" 0.670916 \n",
" -0.103986 \n",
" 0.030084 \n",
" 8.95 \n",
" \n",
" \n",
" 8 \n",
" 133944.0 \n",
" 2.119362 \n",
" 0.142639 \n",
" -2.373337 \n",
" 0.541949 \n",
" 0.608419 \n",
" -1.775564 \n",
" 0.955775 \n",
" -0.599383 \n",
" 0.010420 \n",
" 0.295305 \n",
" -0.936569 \n",
" -0.452478 \n",
" -1.340798 \n",
" 1.077459 \n",
" -0.099584 \n",
" -0.815072 \n",
" 0.018481 \n",
" -0.639446 \n",
" -0.065427 \n",
" -0.323573 \n",
" 0.264264 \n",
" 0.898266 \n",
" -0.168063 \n",
" 0.059311 \n",
" 0.626949 \n",
" 0.729035 \n",
" -0.129120 \n",
" -0.094713 \n",
" 10.00 \n",
" \n",
" \n",
" 9 \n",
" 58769.0 \n",
" -5.584256 \n",
" -4.732413 \n",
" -0.448452 \n",
" -0.121442 \n",
" -0.707412 \n",
" -0.114376 \n",
" -1.554628 \n",
" 1.402126 \n",
" -0.031693 \n",
" -0.942358 \n",
" -2.439501 \n",
" -0.552312 \n",
" -0.295588 \n",
" -0.250246 \n",
" -1.197732 \n",
" 1.549553 \n",
" 0.933237 \n",
" -1.237689 \n",
" 0.416832 \n",
" -1.046900 \n",
" 0.041651 \n",
" 0.621789 \n",
" 0.223467 \n",
" -0.770137 \n",
" 0.621182 \n",
" -0.028738 \n",
" 0.505194 \n",
" -1.898323 \n",
" 101.49 \n",
" \n",
" \n",
"
\n",
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plot_confusion_matrix(y_test, afd_preds)"
]
},
{
"cell_type": "code",
"execution_count": 206,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" precision recall f1-score support\n",
"\n",
" non-fraud 1.00 0.99 0.99 28435\n",
" fraud 0.09 0.91 0.17 46\n",
"\n",
" accuracy 0.99 28481\n",
" macro avg 0.55 0.95 0.58 28481\n",
"weighted avg 1.00 0.99 0.99 28481\n",
"\n"
]
}
],
"source": [
"print(classification_report(\n",
" y_test, afd_preds, target_names=['non-fraud', 'fraud']))"
]
},
{
"cell_type": "code",
"execution_count": 112,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" time \n",
" va \n",
" vb \n",
" vc \n",
" vd \n",
" ve \n",
" vf \n",
" vg \n",
" vh \n",
" vi \n",
" vj \n",
" vk \n",
" vl \n",
" vm \n",
" vn \n",
" vo \n",
" vp \n",
" vq \n",
" vr \n",
" vs \n",
" vt \n",
" vu \n",
" vv \n",
" vw \n",
" vx \n",
" vy \n",
" vz \n",
" vaa \n",
" vab \n",
" amount \n",
" score \n",
" outcomes \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 41505.0 \n",
" -16.5265065691231 \n",
" 8.58497179585822 \n",
" -18.649853185194498 \n",
" 9.50559351508723 \n",
" -13.793818527095699 \n",
" -2.8324042993974703 \n",
" -16.701694296045 \n",
" 7.517343903709871 \n",
" -8.50705863675898 \n",
" -14.110184441545696 \n",
" 5.29923634963938 \n",
" -10.8340064814734 \n",
" 1.67112025332681 \n",
" -9.37385858364976 \n",
" 0.36080564163161705 \n",
" -9.899246540806661 \n",
" -19.2362923697613 \n",
" -8.39855199494575 \n",
" 3.10173536885404 \n",
" -1.51492343527852 \n",
" 1.19073869481428 \n",
" -1.1276700090206102 \n",
" -2.3585787697881 \n",
" 0.6734613289872371 \n",
" -1.4136996745881998 \n",
" -0.46276236139933 \n",
" -2.01857524875161 \n",
" -1.04280416970881 \n",
" 364.19 \n",
" 1000.0 \n",
" [fraud] \n",
" \n",
" \n",
" 16 \n",
" 52913.0 \n",
" 0.676854829386595 \n",
" 1.58702920998622 \n",
" -1.50814698254807 \n",
" 1.44381484435084 \n",
" 1.31678997910859 \n",
" -1.1603421410518702 \n",
" 1.1490491098159 \n",
" -0.4193931328245341 \n",
" -0.0905654487115674 \n",
" -0.323058484892761 \n",
" 2.967920897973221 \n",
" -0.155026280352087 \n",
" -0.7177661091303009 \n",
" -3.70927517991521 \n",
" 0.731050307180359 \n",
" 0.51012502005825 \n",
" 2.7302621301168104 \n",
" 1.46374586218951 \n",
" -0.578405585881194 \n",
" 0.31792704392540505 \n",
" -0.191418401886823 \n",
" 0.101450781492372 \n",
" 0.05829839657805901 \n",
" -0.135311156556857 \n",
" -0.6044061673895129 \n",
" -0.4527632157702071 \n",
" 0.0464301769729269 \n",
" -0.3836415178705061 \n",
" 0.89 \n",
" 977.0 \n",
" [fraud] \n",
" \n",
" \n",
" 19 \n",
" 55575.0 \n",
" -3.41942263819263 \n",
" -1.13007559809015 \n",
" 1.60708637375377 \n",
" 2.2299789220985096 \n",
" 2.64408903133496 \n",
" -0.7935377223504201 \n",
" 1.08824220975816 \n",
" -1.00840988415033 \n",
" -0.933315168258216 \n",
" 1.92198175926539 \n",
" 1.9544453994250897 \n",
" 0.388839766922148 \n",
" 0.30147367767299 \n",
" -0.427868493466856 \n",
" -0.0292266721907548 \n",
" -0.0750638320495107 \n",
" -0.7991270792650841 \n",
" -0.629796480633153 \n",
" -0.469016593690311 \n",
" -1.44533223472583 \n",
" -0.479674205429058 \n",
" 0.524022264453329 \n",
" 0.093685026163671 \n",
" 0.30523015264387393 \n",
" 0.6539626403904271 \n",
" 0.0228521071220258 \n",
" -0.955552398442486 \n",
" -0.0856295445716383 \n",
" 22.7 \n",
" 953.0 \n",
" [fraud] \n",
" \n",
" \n",
" 32 \n",
" 141655.0 \n",
" -29.942972270707894 \n",
" -25.8317821051362 \n",
" -16.2275119291273 \n",
" 6.690679169572508 \n",
" -20.7878458115363 \n",
" 13.0856943709389 \n",
" 17.2566228959585 \n",
" -9.16174621113856 \n",
" 5.00304130464362 \n",
" -2.4314663912742 \n",
" 1.7599893933563102 \n",
" 1.86143271752851 \n",
" 3.05338368623288 \n",
" -1.4096397429751502 \n",
" 3.68027152852471 \n",
" 4.53127503027372 \n",
" 0.22001123266623696 \n",
" -1.60453617589942 \n",
" 3.14031061569242 \n",
" -21.3871219290332 \n",
" -2.49469918146669 \n",
" -0.660297208514391 \n",
" -8.53781580651041 \n",
" 0.40080447201294 \n",
" -0.643022787031134 \n",
" 0.4969028458312689 \n",
" 6.26770908866261 \n",
" -2.7650700017587595 \n",
" 3502.11 \n",
" 962.0 \n",
" [fraud] \n",
" \n",
" \n",
" 51 \n",
" 56803.0 \n",
" 1.17360912432855 \n",
" 0.511603445630379 \n",
" -0.5204514757086011 \n",
" 1.12262620608436 \n",
" 0.35886736932159297 \n",
" -0.29596043353185303 \n",
" 0.0250046387752933 \n",
" 0.14444171501870098 \n",
" -0.110183833896483 \n",
" -0.513358438415108 \n",
" 1.50090553458987 \n",
" -0.228726450784267 \n",
" -1.5703348271985802 \n",
" -0.839677563070961 \n",
" 0.8255113303861942 \n",
" 0.6647743241018189 \n",
" 0.886274830361083 \n",
" 0.925736187467082 \n",
" -0.5606318507221589 \n",
" -0.20184199976003606 \n",
" -0.0268349064777576 \n",
" -0.0755364811334122 \n",
" -0.129545704955994 \n",
" -0.4816867297130711 \n",
" 0.582500674589865 \n",
" -0.289875901401972 \n",
" 0.0379128239034746 \n",
" 0.0374365751819376 \n",
" 1.0 \n",
" 958.0 \n",
" [fraud] \n",
" \n",
" \n",
" 59 \n",
" 94330.0 \n",
" -0.159188253451539 \n",
" 1.6364222543563205 \n",
" 1.54599038025481 \n",
" 4.55479940392864 \n",
" 0.972500161314345 \n",
" 0.135619469003896 \n",
" 0.890709635187871 \n",
" -0.308928981901544 \n",
" -0.750671002033175 \n",
" 1.48808516225536 \n",
" 0.0189227146632523 \n",
" -3.51495112016203 \n",
" 1.37852083915781 \n",
" 1.7509347239023998 \n",
" 0.0364873683299348 \n",
" -0.332543453736555 \n",
" 0.8312407987466272 \n",
" 0.0272426739399668 \n",
" 1.11755527782445 \n",
" 0.374095700292396 \n",
" -0.4149084993816161 \n",
" -0.7253325636219929 \n",
" 0.0198138142745913 \n",
" -0.0804172850283939 \n",
" -0.5666438490691871 \n",
" 0.0294988448472911 \n",
" 0.206860492083663 \n",
" 0.00306000773670604 \n",
" 12.74 \n",
" 965.0 \n",
" [fraud] \n",
" \n",
" \n",
" 177 \n",
" 58954.0 \n",
" 1.23804261799844 \n",
" 0.77053043333537 \n",
" -0.331997523908519 \n",
" 1.34546451219451 \n",
" 0.139816295211697 \n",
" -1.22967596595336 \n",
" 0.354259165638368 \n",
" -0.280804576159513 \n",
" -0.184996149214632 \n",
" -0.699812889267707 \n",
" 0.230655066658352 \n",
" 0.21477524724131605 \n",
" 0.561439138721415 \n",
" -1.4597942610637098 \n",
" 1.0212476623650102 \n",
" 0.567113655972287 \n",
" 0.859052402876907 \n",
" 0.516214961212954 \n",
" -0.629783953300129 \n",
" -0.0582422755483618 \n",
" -0.0605703154557262 \n",
" -0.0710964980063678 \n",
" -0.131000023986718 \n",
" 0.269144247106671 \n",
" 0.7120999799201929 \n",
" -0.32948917861264804 \n",
" 0.039308026136455 \n",
" 0.0588378743714698 \n",
" 1.0 \n",
" 947.0 \n",
" [fraud] \n",
" \n",
" \n",
" 253 \n",
" 878.0 \n",
" -2.5121924071651502 \n",
" 0.31783823871856204 \n",
" 1.27033280212107 \n",
" 1.21425795018979 \n",
" 1.17990250727029 \n",
" -0.41223326510912106 \n",
" 1.00271608138525 \n",
" -0.7746801839401241 \n",
" 1.65660199646692 \n",
" 2.7790737156669403 \n",
" -0.0847781328647432 \n",
" -0.18321933511528 \n",
" -0.330652014184725 \n",
" -1.2773505362658202 \n",
" 0.586282642874345 \n",
" -1.56966184767482 \n",
" 0.122172327014581 \n",
" -0.9167298664751452 \n",
" 0.6695246342509189 \n",
" 0.5929484938755051 \n",
" -0.655257798829788 \n",
" 0.24236823281838 \n",
" -0.135073900073597 \n",
" 0.104260701044715 \n",
" 0.211102067974728 \n",
" -0.28730667236331103 \n",
" 0.335129131058453 \n",
" 0.32821181910776 \n",
" 18.8 \n",
" 962.0 \n",
" [fraud] \n",
" \n",
" \n",
" 328 \n",
" 110199.0 \n",
" 1.75136324325772 \n",
" 0.337812400833952 \n",
" 0.795262392956418 \n",
" 4.31273176972153 \n",
" -0.681348924648448 \n",
" -0.126233056366034 \n",
" -0.566591255980354 \n",
" -0.0313440579660118 \n",
" 0.919522880708254 \n",
" 0.975130081266523 \n",
" 0.435764202447928 \n",
" -2.21569585920562 \n",
" 1.91457117907994 \n",
" 1.15256247746996 \n",
" -1.3746570764088202 \n",
" 0.8239809240929471 \n",
" 0.175328666447042 \n",
" -0.0483307755538258 \n",
" -2.38942182862032 \n",
" -0.289943444109214 \n",
" 0.16007070110616595 \n",
" 0.803841115515162 \n",
" 0.25356928092348496 \n",
" 0.8074573633362009 \n",
" -0.3400690874520441 \n",
" -0.00655722686278409 \n",
" -0.00813323923492112 \n",
" -0.0250715519401937 \n",
" 29.58 \n",
" 955.0 \n",
" [fraud] \n",
" \n",
" \n",
" 352 \n",
" 164441.0 \n",
" 1.88776142403002 \n",
" 0.62826988690252 \n",
" -2.75957058286354 \n",
" 1.62239582744255 \n",
" 0.9446222968645832 \n",
" -1.15455901266687 \n",
" 0.364580675536746 \n",
" -0.13538934616727902 \n",
" 0.0345521682776357 \n",
" -0.9746950773303552 \n",
" 1.53126492421627 \n",
" -0.558651257008026 \n",
" -1.40204078796233 \n",
" -2.7255621085966 \n",
" 0.405777909218057 \n",
" 1.43115948321574 \n",
" 1.95612710013755 \n",
" 2.26414005330485 \n",
" -0.8243269352259971 \n",
" -0.13330303807844898 \n",
" 0.106823573857281 \n",
" 0.29362617939436403 \n",
" -0.136295490129905 \n",
" -0.7309626721782321 \n",
" 0.263143555477082 \n",
" -0.4195455936504099 \n",
" 0.0292391204906618 \n",
" 0.0198264451018551 \n",
" 48.8 \n",
" 976.0 \n",
" [fraud] \n",
" \n",
" \n",
" 361 \n",
" 68878.0 \n",
" -4.80458628691832 \n",
" -4.35021922143654 \n",
" 0.8547339441976851 \n",
" 0.207810580386234 \n",
" 2.91117776563597 \n",
" -2.16185200836836 \n",
" -2.10165901071884 \n",
" -1.60972469419277 \n",
" -1.24561305676715 \n",
" 1.13640900411353 \n",
" 1.56282964824949 \n",
" -0.656243196361657 \n",
" -1.35367268021855 \n",
" 0.499765130779605 \n",
" 0.8153495186707951 \n",
" 0.7731699739080921 \n",
" 0.809416192491745 \n",
" -1.4911056392524902 \n",
" 2.03964533568703 \n",
" -2.1686234190058897 \n",
" 0.759812961941168 \n",
" -0.6550949394591179 \n",
" 1.64631867508059 \n",
" 0.0487933593234721 \n",
" 0.8261049032712451 \n",
" -0.0957606816931365 \n",
" 1.3059493867690302 \n",
" -0.579378536086197 \n",
" 60.17 \n",
" 960.0 \n",
" [fraud] \n",
" \n",
" \n",
" 366 \n",
" 23407.0 \n",
" 1.24074673650722 \n",
" 0.848160288821537 \n",
" -0.199824951789007 \n",
" 1.55372383755802 \n",
" 0.171272503594932 \n",
" -1.19411648350141 \n",
" 0.228779262997882 \n",
" -0.333950404622801 \n",
" 1.17298661915574 \n",
" -1.05622233741673 \n",
" 1.09596426457221 \n",
" -2.54864074095523 \n",
" 1.71897660311815 \n",
" 0.0708382663905947 \n",
" 0.33670269233951194 \n",
" 0.56048474390029 \n",
" 1.6189262744826105 \n",
" 0.8487676121649308 \n",
" -0.8297510609350229 \n",
" -0.13240918359282802 \n",
" -0.216617664690264 \n",
" -0.2968326218353089 \n",
" -0.113442195893096 \n",
" 0.194913011896955 \n",
" 0.67266243405551 \n",
" -0.366471839903345 \n",
" 0.0094225142703802 \n",
" 0.0513888538008156 \n",
" 1.0 \n",
" 902.0 \n",
" [fraud] \n",
" \n",
" \n",
" 467 \n",
" 53370.0 \n",
" -2.52338699856657 \n",
" -5.90958900863933 \n",
" -2.1145490761038404 \n",
" 2.68771715054533 \n",
" -1.7189459546685002 \n",
" 0.7883472392992701 \n",
" 2.62144747179771 \n",
" -0.387473414390844 \n",
" -0.304782617317214 \n",
" -0.7501911961773671 \n",
" 0.8925662599842149 \n",
" 0.679212965658016 \n",
" -0.546673452270584 \n",
" 1.03001531809147 \n",
" -0.116777526168877 \n",
" -0.09147870610096 \n",
" -0.10583625004301 \n",
" 0.193687825429198 \n",
" -0.682010917016811 \n",
" 3.67662607211696 \n",
" 1.17496005535309 \n",
" -0.8406424904120059 \n",
" -1.85015434331846 \n",
" -0.21348991784278 \n",
" 0.0526663052725845 \n",
" -0.42989130448891705 \n",
" -0.344993893936107 \n",
" 0.328742544686591 \n",
" 1918.5 \n",
" 907.0 \n",
" [fraud] \n",
" \n",
" \n",
"
\n",
"