0

I am working on a project in which I have about 18 classes with about 4,000 total instances. I have 7 attributes, 1 being string data, the rest nominal. I am currently using StringToWordVector on the string attribute with Platt's SMO classifier, achieving good results. We are about to implement this, but I would like to try other classifiers in case there maybe one I could get better results from. Any suggestions?

Also, should I be using MultiClassClassifier with so many classes? If so, what settings should I try within that?

Any advice is appreciated!

Will Ebert
  • 21
  • 6
  • Try multinomial Naive bayes classifier. It is preferred text classification for It gives results at par with SVM. It will be faster too. – Rahul Rajput May 19 '16 at 10:18
  • Multinomial Naive Bayes is not compatible with multi-valued nominal attributes is the error I receive when I try to run the model. Naive bayes performed with about 20% less accuracy than SMO. I haven't tried all of the possibilities with MuliClassClassifier and Naive bayes yet but I don't see it performing much better. Thanks for the recommendation though! – Will Ebert May 20 '16 at 19:37

1 Answers1

0

An AdaBoosted J48 Decision Tree yielded the best results has been well established in our division

Will Ebert
  • 21
  • 6