Machine Learning
- Split train, validation, and test data before tuning to prevent leakage.
- Choose metrics for the actual cost of false positives and false negatives.
- Monitor data drift and real-world outcomes after deployment.
Cool NotesML evaluation essentials
All notesML evaluation essentials
Machine Learning · Original AlgoGym revision note with a source for deeper study.