feature selection by using random forest
所属分类:界面编程
开发工具:Python
文件大小:28939KB
下载次数:2
上传日期:2020-12-27 15:21:42
上 传 者:
tendless00
说明: 随机森林做特征选择py程序feature selection by using random forest
(feature selection by using random forest)
文件列表:
feature_selection (0, 2017-04-01)
feature_selection\Dataset (0, 2017-04-01)
feature_selection\Dataset\LargeTrain.csv (90797153, 2017-04-01)
feature_selection\RF_feature_selection.py (1397, 2017-04-01)
feature_selection\RF_top10.csv (506870, 2017-04-01)
feature_selection\requirements.in (69, 2017-04-01)
feature_selection\requirements.txt (588, 2017-04-01)
# Feature_selection
use random forest to find important feature.
## 1. feature importance ranking:
* ### By Feature importance ranking top10
+ 1. TB_9a
+ 2. TB_a9
+ 3. Img0.1
+ 4. ent_p_5
+ 5. TB_b1
+ 6. TB_71
+ 7. ent_p_8
+ 8. TB_ce
+ 9. GetStringTypeA
+ 10. ExitProcess
## 2. Useless feature ranking:
* ### Featurn ranking list Top10 useless feature:
+ GetLastActivePopup
+ ImageList_Add
+ GlobalDeleteAtom
+ IsBadReadPtr
+ SelectPalette
+ GetMenuState
+ ExitThread
+ AdjustWindowRectEx
+ GetEnvironmentVariableA
+ SHGetFileInfoA
## 3. decision method
+ 使用 sklearn 裡面的 RandonforestClassifier,取前10為 important features,後10為useless features。
+ useless features 只代表**相對不佳**,不代表這些 useless features 對分類**完全沒用**,important features 亦同。
## 4. packages
+ sklearn package 中的 RandomForestClassifier 做 feature selection
+ pandas,numpy 處理資料
+ matplotlib.pyplot 畫圖
## 5. suggestion
+ 無
近期下载者:
相关文件:
收藏者: