Options

"Using SVM for regression"

Phillip_KoshutePhillip_Koshute Member Posts: 2 Contributor I
edited May 2019 in Help
How should the parameters for the SVM process be set so that it performs regression rather than classification? Thanks!
Tagged:

Answers

  • Options
    landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    this depends on the SVM  implementation you use. The LibSVM needs to be switched to one of the SVR Modes (first parameter). The standard SVM does this automatically.

    Greetings,
      Sebastian
  • Options
    Phillip_KoshutePhillip_Koshute Member Posts: 2 Contributor I
    Thanks Sebastian,

    I think that I am gaining a better understanding of how SVM may be used for regression. In RapidMiner, I am using the standard SVM operator. I have noticed for the different kernels that there are a variety of parameters. However, it does not seem as though there is much explanation of the significance of each parameter in the RapidMiner Wiki. If possible, could you please point me to any resource that would explain more about these parameters? Alternatively, can you offer a summary of the different parameters?

    Thanks very much,
        Phil
  • Options
    landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    these parameters are very much SVM standard. Please refer to any appropriate literature for that, I don't have one right one. You are very invited to share your gained knowledge in the wiki, leaving it for others...

    Greetings,
      Sebastian
Sign In or Register to comment.