Сredit scoring competition based on credit history data from the Alfa-Bank Machine Learning Laboratory and the Open Data Science Community.
Credits for the training sample are taken over a period of M months, and credits for the test sample are taken over the next K months.
Whole train set shape: 3000000, 450;
Whole test set shape: 500000, 450.
The target variable is a binary value that takes the values 0 and 1, where 1 corresponds to the client's default on the loan.
Competition metric is ROC AUC.
Public leaderboard: 0.751802;
Private leaderboard: 0.748337.