Options

"Weighted SVM"

m_r_nourm_r_nour Member Posts: 35 Maven
edited May 2019 in Help
Weighted SVM is an excellent classifier to deal with unbalanced data.

Do you have any idea , how we can use it in RM?


Tagged:

Answers

  • Options
    RalfKlinkenbergRalfKlinkenberg Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, RMResearcher, Member, Unconfirmed, University Professor Posts: 68 RM Founder
    RapidMiner provides several SVM implementations (RapidMiner operators JMySVM, LibSVM, EvoSVM, PsoSVM, ...) and several ways to use weights and/or cost factors:
       
    • JMySVM supports class weights for two class problems.
      •  
    • MetaCost can be used to assign different misclassification costs (weights) to different classes.
  • Options
    m_r_nourm_r_nour Member Posts: 35 Maven
    thank you


    but I do not know how to use Libsvm and moreover how to use meta cost  to tune parameters of Libsvm
    ratio  of no. of data in my dataset are:
    50%,25%,10%,7%,5%,2%,1%
    no of sample 3120 and number of features :57

    I'd be appreciate if you help me to deal with this problem

    regards
    REZA
Sign In or Register to comment.