search

LenzDu / Kaggle-Competition-Favorita

5th place solution for Kaggle competition Favorita Grocery Sales Forecasting
MIT License
250 stars 96 forks source link

statistical feature issue from lgbm.py #5

Open purezhanghan opened 5 years ago

purezhanghan commented 5 years ago

Firstly thanks for this great solution for the competition. I found the following lines from lgbm.py script[line 52] and the statistical feature median value is actually implemented with mean method, is that a typo? X['median_{}_2017'.format(i)] = get_timespan(df_2017, t2017, i, i).mean(axis=1).values

LenzDu commented 5 years ago

It's a typo, we got two identical features accidentally ;)