Examples of weka.filters.unsupervised.attribute.PrincipalComponents

weka.filters.unsupervised.attribute.PrincipalComponents

Performs a principal components analysis and transformation of the data.
Dimensionality reduction is accomplished by choosing enough eigenvectors to account for some percentage of the variance in the original data -- default 0.95 (95%).
Based on code of the attribute selection scheme 'PrincipalComponents' by Mark Hall and Gabi Schmidberger.

Valid options are:

 -D Don't normalize input data.

 -R <num> Retain enough PC attributes to account for this proportion of variance in the original data. (default: 0.95)

 -A <num> Maximum number of attributes to include in  transformed attribute names. (-1 = include all, default: 5)

 -M <num> Maximum number of PC attributes to retain. (-1 = include all, default: -1)

@author Mark Hall (mhall@cs.waikato.ac.nz) -- attribute selection code @author Gabi Schmidberger (gabi@cs.waikato.ac.nz) -- attribute selection code @author fracpete (fracpete at waikato dot ac dot nz) -- filter code @version $Revision: 6690 $


  /**
   * Default projection method.
   */
  protected Filter defaultFilter() {
    PrincipalComponents filter = new PrincipalComponents();
    //filter.setNormalize(false);
    filter.setVarianceCovered(1.0);
    return filter;
  }

        
        return filter;
    }
    
    protected Filter getFeatureReductionFilter(ExampleSet examples) throws Exception {
        PrincipalComponents filter = new PrincipalComponents();
        filter.setMaximumAttributes(10);
        
        filter.setInputFormat(examples.getInstances());
        Instances filtered = Filter.useFilter(examples.getInstances(), filter);
        examples.setInstances(filtered);
        
        return filter;
    }

Examples of weka.filters.unsupervised.attribute.PrincipalComponents

Related Classes of weka.filters.unsupervised.attribute.PrincipalComponents