JDD-SalesPredictionChallenge

JDD-Cup-2017-SessionA.
An open source baseline, but it is not a good way to deal with this compition.
If you have any problem, send them to my Email: cs_xcy@126.com
best online score by this code: 0.4804

static_*.py:

This code can get the feature in a long time, it responses a stable factor of data.

*_all.py:

This can response a temporary feature.

*_cv.py:

It is to make cross vilidation.

xgboost_*.py

Made model.