Following this, We saw Shanth’s kernel on starting new features from the `bureau
Element Technology
csv` dining table, and i began to Google several things such as for example “Tips profit an excellent Kaggle battle”. Most of the results said that the secret to profitable are feature engineering. Therefore, I decided to element professional, but since i did not actually know Python I will not do they to your hand regarding Oliver, therefore i returned so you’re able to kxx’s code. I function engineered some posts considering Shanth’s kernel (We hand-blogged aside all groups. ) following fed they toward xgboost. It got regional Curriculum vitae away from 0.772, along with personal Lb away from 0.768 and private Lb of 0.773. Therefore, my personal element systems did not help. Awful! Thus far I was not therefore reliable out-of xgboost, thus i attempted to rewrite this new code to utilize `glmnet` having fun with collection `caret`, but I didn’t understand how to enhance a blunder I had while using the `tidyverse`, therefore i averted. You can observe my personal code of the pressing here. More