Enhanced AML fraud detection solutions with Azure Machine Learning
summary
A working AML pipeline uses Azure Machine Learning Studio to merge client and transaction feeds, engineers features (including a 89% credit-card skew), trains a multi-class decision forest on ~834,000 training rows with an 80/20 split, scores and compares multiple models in one experiment, and deploys a predictive web service with URIs and sample code for real-time or batch scoring. Report_flag serves as the label, the demo claims near-perfect fit on the sample but emphasizes iterative retraining and false-positive reduction.
AI-written summary · about
chapters (28)
- 0:00Introduction
- 0:38What is AML
- 1:01What is Money Laundering
- 2:52Typical AML Flow
- 4:09Common AML Pitfalls
- 5:52Enhanced AML Fraud Detection
- 7:49Machine Learning for AML
- 8:49Machine Learning with Azure
- 14:18Azure Machine Learning Studio
- 14:56Machine Learning Studio
- 20:45Data Visualization
- 22:53Machine Learning
- 24:13Train Model
- 25:13Running the Experiment
- 27:44Evaluate Model
- 28:30Web Service Setup
- 28:47Web Service Output
- 29:43Deploy Web Service
- 30:09Multiple Machine Learning Models
- 31:36Demo
- 34:34Case management
- 37:26Insurance fraud
- 38:35Visual workflow development
- 39:58Data Jet
- 40:33Report Column
- 42:15Straind Model
- 46:52Retraining Pipeline
- 47:32Be Testing
related talks
46mML for Anti-Money Laundering
01h 05mAzure Machine Learning for the Developer
41mLeveraging Automated Machine Learning to enable anyone to develop Machine Learning solutions
23mChallenges in using machine learning for financial monitoring
51mTaking Models to the Next Level with Azure Machine Learning
39mFraud Detection: Using ML to Identify and Manage Fraudulent Activities
58mDeep Dive into Azure Automated Machine Learning
16mLeveraging ML to Create a Reliable Anti-Fraud System