Skip to content

Comment on Kaggle Post-Mortem: The dangers of overfitting

Comments

Why even look at these scoreboards?

Assume that the validation data is similar to the test data. Then, by performing experiments (making submissions) and recording measurements (analysing the returned leaderboard score) you can attempt to infer something about the structure of the validation data, and make more accurate predictions about the test data. You can probably infer quite a bit of useful information by looking at other people's scores too, even if you don't have access to their submitted predictions.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.