creke / rubbishmessage Goto Github PK

w.r.t by 豌豆先生

In folder src includes three file that can run independently, 
first ,run >>python keyWords.py to get classLabel.txt and trainLeft.txt ...
then run >>python crossTest.py ,result is
TDIDF+NB: F = 95.5
TDIDF+SVM(LinearSVC): F = 97.5
TDIDF+RF: F = 97.5 
Counter+LinearSVC: F = 98.2
Hash+TDIDF+LinearSVC: F = 97.8
...
then run >> python IG.py,use IG(information gain),or MI(multiple information),or WLLR
result is:
MI: F = 97.5
IG,WLLR: F = 98.5


LDA is testing,look forward to  the result