主要是修改Kernelhttps://www.kaggle.com/gaborfodor/from-eda-to-the-top-lb-0-367
Stacking:https://github.com/freelzy/Tencent_Social_Ads
修改部分:调参,Stacking,线性加权
pre_processing.py:特征工程
stacking_processing.py:xgb和lgb5折cv对训练集和测试集分别预测,作为两列特征保存
Stacking_lgb.py:把xgb 5折cv生成的一列特征加到原始特征上,用lgb进行5折cv
Stacking_xgb.py:把lgb 5折cv生成的一列特征加到原始特征上,用xgb进行5折cv
Line_stacking.py:线性加权
模型:Xgboost,lightGBM
最优成绩:Public 0.37210,(xgb 5折cv)*0.6+(lgb stacking)*0.4
说明:最优成绩并不是xgb_stacking与lgb_stacking线性加权得到,而是xgb 5折cv与lgb_stacking线性加权得到
目的:保存代码
致谢:感谢Kernel部分各位选手的开源方案,受益匪浅
-
Notifications
You must be signed in to change notification settings - Fork 3
LightR0/New-York-City-Taxi-Trip-Duration
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
Kaggle比赛New-York-City-Taxi-Trip-Duration方案
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published