Expresso Churn Prediction Challenge
'%3e%3cg id='b'%3e%3cpath id='a' d='M0-100V0h50z' transform='rotate(18 0 -100)'/%3e%3cuse xlink:href='%23a' transform='scale(-1 1)'/%3e%3c/g%3e%3cuse xlink:href='%23b' transform='rotate(72)'/%3e%3cuse xlink:href='%23b' transform='rotate(144)'/%3e%3cuse xlink:href='%23b' transform='rotate(216)'/%3e%3cuse xlink:href='%23b' transform='rotate(288)'/%3e%3c/g%3e%3c/svg%3e)
Helping Senegal
$1 000 USD
Completed (over 4 years ago)
Classification
Prediction
1378 joined
437 active
Start
Aug 27, 21
Close
Nov 28, 21
Reveal
Nov 28, 21
0.0+ score on leaderboard
Help · 19 Oct 2021, 19:25 · 9
My first and second submissions gave a roc_auc of 0.0+ score on leaderboard using lgbm
But while training and cross validation on Google Colab I recorded a training roc_auc socre of 0.9+ and CV roc_score score of 0.9+
Can someone please help me out?
Make sure your submission file contains probabilities between 0 and 1 inclusive.
i made use of predict_proba but still same result on leaderboard
Are you getting any errors while working or it's simply the score that is not showing properly?
It is simply the scores that is not showing up properly
Hello, do you run predict_proba[:, 1] or this one [:,0] for your submission?
I ran only predict_proba(test)
Ok, predict proba gives an array of shape no of rows X 2 columns. The 1st column represents probability of 0, 2nd probability of 1(which in our case is probability of Churn needed for an AUC metric).
Run predict_proba(test)[:, 1] and see if it fixes your situation
Thank you so much,
It fixed it
Cheers!