Chess Anomaly Detection: Machine learning
Lichess games become player-level records, six unsupervised detectors are compared against synthetic cheats, and an ensemble flagged 312 of 17,909 players for human review. It scored 100/100 as the ML final project.