online payment fraud detection with azure machine learning

13
Online Payment Fraud Detection with Azure Machine Learning

Upload: stefano-tempesta

Post on 22-Jan-2018

58 views

Category:

Software


0 download

TRANSCRIPT

Page 1: Online Payment Fraud Detection with Azure Machine Learning

Online Payment Fraud Detectionwith Azure Machine Learning

Page 2: Online Payment Fraud Detection with Azure Machine Learning

Stefano Tempesta

Page 3: Online Payment Fraud Detection with Azure Machine Learning

AGENDA

Intro to Azure Machine Learning

Payment Fraud Detection

Anomaly Detection in Azure ML

Page 4: Online Payment Fraud Detection with Azure Machine Learning
Page 5: Online Payment Fraud Detection with Azure Machine Learning

Azure Machine Learning

• Sentiment Analysis

• Demand Estimation

• Recommendations

• Outcome Prediction

MODEL REST ServiceDATASET

EXPERIMENT

• Training

• Scoring

• Evaluation

Page 6: Online Payment Fraud Detection with Azure Machine Learning
Page 7: Online Payment Fraud Detection with Azure Machine Learning

VISA handles

2000 transactions / sec

>170M tpd

In 2015, financial fraud totaled a cost of€ 900 million

Page 8: Online Payment Fraud Detection with Azure Machine Learning

Traditional Payment Fraud Detection

• Analyze transactions and human-review suspicious ones

• Use a combination of data, horizon-scanning and “gut-feel”

• Every attempted purchase that raises an alert is either declined or reviewed• False positive

• Fail to predict unware threats

• Need to update risk score regularly

Page 9: Online Payment Fraud Detection with Azure Machine Learning

ML Payment Fraud Detection

• Large, historical datasets across many clients and industries• Benefit also small companies

• Self-learning models not determined by a fraud analyst• Update risk score quickly

Page 10: Online Payment Fraud Detection with Azure Machine Learning

Feature Engineering

• Aggregated variables: aggregated transaction amount per account, aggregated transaction count per account in last 24 hours and last 30 days.

• Mismatch values: mismatch between shipping Country and billing Country.

• Risk tables: fraud risks are calculated using historical probability grouped by country, IP address, etc.

Page 11: Online Payment Fraud Detection with Azure Machine Learning

Anomaly Detection

• Anomaly Detection encompasses many important tasks in Machine Learning:• Identifying transactions that are potentially fraudulent

• Learning patterns that indicate an intrusion has occurred

• Finding abnormal clusters of patients

• Checking values input to a system

• Azure Machine Learning supports:• PCA-Based Anomaly Detection Principal Components

• One-Class Support Vector Machine Supervised Learning

Page 12: Online Payment Fraud Detection with Azure Machine Learning

Anomaly Detection in Azure ML

Page 13: Online Payment Fraud Detection with Azure Machine Learning