목차
Approaches for target metric optimization
Early stopping
Conclusion
관련 문서
Approaches for target metric optimization
Optimize another metric, postprocess
Write custom loss function
Any, if you can
Early stopping
Optimize metric M1, monitor metric M2
Stop when M2 score is the best
Conclusion
Loss vs metric
Approaches in general:
Just run the right model
관련 문서
How to Win a Data Science Competition: Learn from Top Kagglers