Zindi User Behaviour Birthday Challenge
Helping Africa
$3 000 USD
Completed (almost 5 years ago)
Prediction
882 joined
174 active
Start
24 Sep 21
Close
23 Jan 22
Reveal
23 Jan 22
'%3e%3ccircle r='20' fill='%23008'/%3e%3ccircle r='17.5' fill='%23fff'/%3e%3ccircle r='3.5' fill='%23008'/%3e%3cg id='d'%3e%3cg id='c'%3e%3cg id='b'%3e%3cg id='a' fill='%23008'%3e%3ccircle r='.875' transform='rotate(7.5 -8.75 133.5)'/%3e%3cpath d='M0 17.5.6 7 0 2l-.6 5L0 17.5z'/%3e%3c/g%3e%3cuse xlink:href='%23a' transform='rotate(15)'/%3e%3c/g%3e%3cuse xlink:href='%23b' transform='rotate(30)'/%3e%3c/g%3e%3cuse xlink:href='%23c' transform='rotate(60)'/%3e%3c/g%3e%3cuse xlink:href='%23d' transform='rotate(120)'/%3e%3cuse xlink:href='%23d' transform='rotate(-120)'/%3e%3c/g%3e%3c/svg%3e)
Difference in Validation and LB auc
Data · 27 Dec 2021, 09:41 · 8
My 5 fold CV gives AUC around 92% with both Neural Net as well as Tree Boosting, However the LB auc is below 75% and I couldn't figure out why that's the case.
Anyone experienced something similar ?
Sometimes it's because of unbalanced
What's the validation method you are using ? The reported auc will be quite different based on how you are validating
I am validating 5 fold model and on the final Target, not on the individual sub targets. Checked my pipeline as well for any inconsistencies, and the validation results look all okay as expected.
Hope to find the issue soon.
all the best
How did you calculate the AUC locally?
validating 5 fold model on the final Target
@tarickMorty How did you solve the issue? I am experiencing the same problem.
My test dataset preparation had some inconsistency.