My first Kaggle competition

Last month I entered my first Kaggle competition, working with BNP Paribas Cardif’s data. The official aim was to speed up the claims management process; my own goal was simply to put some machine learning techniques into practice. That officially makes me a Kaggler. I used the xgboost R package to implement gradient boosting. Now the results are in, it’s clear I’ve got plenty of room to improve — next time I’ll need to put more work into feature engineering and ensembling my models.