I wanna increase the accuracy of my classification model. What should I do?


I've built a workflow on a classification problem using the decision tree algorithm. What is your solution for increasing the model accuracy using this algorithm. To be honest, I'm the beginer.




this is more a predictive analytics problem not necessarily a text mining problem. If you are new to predictive analytics I suggest to read a book or tutorials on the web.

In general you could do:

  • Different model learnes (SVN)
  • Ensembles (Boosting, Bagging)
  • Parameter Optimization
  • Feature Selection
  • NGram Features

Cheers, Kilian