Why Fraud Detection, Sports Betting, and Bitcoin Prediction Look Alike
summary
Claims a unified ML lifecycle across fraud detection, sports betting, and Bitcoin prediction and gives concrete constraints: payments require 30 ms latency and delayed labels, Bitcoin prediction asks binary up/down decisions with chosen time horizons, sports betting faces tiny sample sizes and must optimize expected value. Offers practical results: insist on a successful backtest and forward test, warns that a complex paper-trading pipeline can have costs that exceed profits and recounts building live betting bots with mixed outcomes.
AI-written summary · about
chapters (6)
- 0:00Fraud, sports, Bitcoin similarities
- 2:50Business and data understanding
- 7:33Model development challenges
- 14:12Model operations and deployment
- 17:18Key takeaways
- 19:44Q&A
related talks
16mLeveraging ML to Create a Reliable Anti-Fraud System
37mThe Role of Machine Learning in Today’s Payment Ecosystems
45mAnti-Fraud in Reverse and the Use of ML Methods in It
13mHarnessing Machine Learning to Combat E-commerce Fraud
40mFraud detection without labeled data
18mDesigning a scalable ML Ops Pipeline: Insights and best practices
48mMachine learning the hard way -- a story about ponies
26mAI-Powered Fraud Detection in Cloud-Native Digital Payments