David Talby, SVP Engineering, Atigeo
summary
Shows how to incrementally build a hybrid fraud-detection model that combines NLP, topic modeling, time-series analysis, link analysis, heuristic rules and anomaly detection. Uses Python with popular open-source data-science libraries and Apache Spark for scalable processing, and demonstrates the approach with a live code walkthrough and public email datasets.
AI-written summary · about
chapters (5)
- 0:00Start
- 2:59Why Semi-Supervised Learning & Feedback?
- 5:14Why Hybrid Analytics?
- 10:54Stream processing
- 12:18User Analysis Iteration
related talks
24mHunting Criminals with Hybrid Analytics
16mLeveraging ML to Create a Reliable Anti-Fraud System
01h 30mAI in Finance From Insight to Action
56mData, AI, Hype and Fear - Recipe for Disaster?
22mReal-world applications of machine learning: Beyond the hype
16mAI-Driven Behavioral Analytics for Real-Time Fraud Prevention
13mTransforming Financial Services with Real-Time Analytics
16mOperationalizing AI in Finance: Real-Time Analytics