1

I was reading a Kaggler Interview article and they kept specifying the importance of a stable and good cross-validation in order to win their competitions. What do they mean by that? I usually just use cross_val_score, and that's enough for me.

Richard Hardy
  • 54,375
  • 10
  • 95
  • 219
Chipmunkafy
  • 115
  • 1
  • 4
  • 3
    Context is everything in these situations. Can you post a link to the article and provide a page number of paragraph if necessary? – StatsStudent Dec 20 '18 at 21:14
  • 1
    @StatsStudent http://blog.kaggle.com/2018/05/07/profiling-top-kagglers-bestfitting-currently-1-in-the-world/ – Chipmunkafy Dec 20 '18 at 22:15
  • Let's put it this way: Do we find that `cross_val_score` gives us good estimates about the performance of our learner in terms of Kaggle leader-board? If, yes, then it is "good". If, no, then it is "bad". (For those who are unaware of how Kaggle works, the leaderboard ranking is usually based on unseen data.) – usεr11852 Dec 21 '18 at 00:23

0 Answers0