Applying a Model on New Dataset.

mrushdi

Hello,

First of all I want to say sorry that im new in this forum, and not familiar with the different options available.
Also I would like to thank Rapid-I team for there patient with us 

My question: I need to test my model that generated by SVM on a new dataset.
First I tranied my dataset using LibSVMLearner and write it to a model. Now I want to apply this model to a new dataset not belong to the training dataset.  Could you please tell me how I can apply it on new dataset. Thanks in advance. 

satyendra

hiiiiiii,
        i hope you are making operator chain that looks like this.if not than try it out.
          root
          |__excel example source
          |
          |_xvalidation
          |            |
          |            |__operator chain
          |            |                |
          |            |                |__libsvm learner                                    //for learning
          |            |                |
          |            |                |__model writer                  //it creates the model file (.mod).save this file.
          |            |
          |            |
          |            |__operator chain2
          |                            |
          |                            |__model applier
          |                            |
          |                            |__bionomialclassification performance    //tells the correctness of created model
          |
          |
          |
          |____excelexample source2      //here you will load the new data set for which you want to do the prediction
          |
          |____model loader                  //here load the (.mod file ) the model
          |
          |____model applier2   

          run this.....and u will get the predictions.
                                with regards,
                                              satyendra

land

Hi,
together with RapidMiner are shipped several sample processes, which will guide you through those basic tasks and should teach you a basic understanding of the concepts. I would recommend to go through all of them.

By the way, satyendra,
I would suggest using the "create_complete_model" option of the XValidation instead of writing each model to file. The problem is, that this way the model you are finally using is not trained on the complete training set but only on the folds currently used in the cross validation.

Greetings,
  Sebastian