0

I am working on a text classification problem in which the 100 most frequent words are selected as features. I believe the results could be improved if I use a better feature selection method? Any ideas? Could TF-IDF work? If yes, then how?

user2295350
  • 303
  • 4
  • 13
  • This question appears to be off-topic because it is about text classification, probably better suited for http://scicomp.stackexchange.com/ – EdChum Oct 07 '13 at 09:03
  • Possible duplicate: https://stackoverflow.com/questions/13603882/feature-selection-and-reduction-for-text-classification?rq=1 – StatguyUser Jun 09 '20 at 04:30

1 Answers1

0

to improve the results you can use Feature selection

1) Information gain

2) Chi square

3) Mutual information

4) Term frequency

TF-IDF you can see this link it will help you

Ashish
  • 1,943
  • 2
  • 14
  • 17