Primary competition visual

Sea Turtle Rescue: Error Detection Challenge

Helping Kenya
$1 950 USD
Challenge completed over 6 years ago
Classification
Anomaly Detection
321 joined
51 active
Starti
Nov 30, 18
Closei
Apr 28, 19
Reveali
Apr 29, 19
Next Topic: Cross-Validation
Help · 13 Mar 2019, 10:05 · 4

Did anyone experience huge discrepancy between local cross-validation score and leaderboard score?

I'm doing 5 folds k-fold crossvalidation, here's what I've been getting in terms of CV vs LB scores: 0.029 -> 0.081

0.027 -> 0.067

0,049 -> 0,058

Discussion 4 answers

I think you mean 0.019 not 0.049 ?

13 Mar 2019, 18:14
Upvotes 0

No typo there: 0.049-something. Correctly building the target variable is (IMO shouldn't be) the first step in solving the problem.

Finding a proper cross-validation process is the next part of the problem. Hint: K-fold cross-validation might not be the fittest here due to the temporal dimension in the dataset.

One thing i forgot is that there might be no correlation between CV and LB score which might make it difficult to build a model that generalises well

Short Message is that its hard to trust your local CV since Zindi competitions always have 50/50 on Public leaderboard, if its was say 30% on Public Leaderboard then you would highly trust your Local CV. Take the message for competition purpose only. thanks

13 Mar 2019, 19:57
Upvotes 0