Skip to content

Latest commit

 

History

History
7 lines (7 loc) · 422 Bytes

README.md

File metadata and controls

7 lines (7 loc) · 422 Bytes

titanic_kaggle

A model classifying whether a person would survive on Titanic
Data: https://www.kaggle.com/c/titanic
The main model is catboost.
Categorial features were encoded with one-hot. To find the best features set I used greedy deletion algorithm.
Also I used 2 grid searches - catboost grid_search and grid search written by me.
Validation result - 0.909(ROC_AUC)