text prediction analysis

lambamanika07lambamanika07 Member Posts: 24 Maven
edited December 2018 in Help

The performance result of text mining model is 79% accuracy model by naive bayes and 83% by SVM (Linear). Is this a god model?

If not can someone please help me to how to improve it. Thank you. A fast response will be appreciated.

Answers

  • lionelderkrikorlionelderkrikor Moderator, RapidMiner Certified Analyst, Member Posts: 1,195 Unicorn

    Hi @lambamanika07,

     

    To improve the performances of your model, you can optimize it using the Optimize Parameters operator and /or 

    try to train other predictive models.

    Pavithra_Rao : Automatic model selection and optimization .

    This tool (to adapt to your own project) select the best model among several optimized models.

     

    And to conclude and to answer to your question "Is this a god model?", I would quote Albert Einstein : "All is relative..."

     

    I hope it will be helpful.

     

    Regards,

     

    Lionel

  • Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

    IF those are your initial results, then I'd say you have a very good model to start with. I would do what @lionelderkrikor suggests and try optimization. You might even want to try some enesmbling methods, like Voting or Stacking to combine the SVM and NB.

Sign In or Register to comment.