first download data from the compeletion website
put three files in inside data folder
cd CTRP_kgaggle;
./unzip_data.sh;
Exploratroy data analysis: EDA_CTRP.ipynb thing might need to be handle
- data ammount is huge.
- too many categorical variable
- lots of variables appear in training but not testing.
Trial script1 lightgbm/xgb sckit learn feature hashing : Building_Models.ipnb Trial script2 using md5 feature hashing with xgb : Building_Models_2.ipnb