Examples of weka.filters.unsupervised.attribute.Reorder

weka.filters.unsupervised.attribute.Reorder

A filter that generates output with a new order of the attributes. Useful if one wants to move an attribute to the end to use it as class attribute (e.g. with using "-R 2-last,1").
But it's not only possible to change the order of all the attributes, but also to leave out attributes. E.g. if you have 10 attributes, you can generate the following output order: 1,3,5,7,9,10 or 10,1-5.
You can also duplicate attributes, e.g. for further processing later on: e.g. 1,1,1,4,4,4,2,2,2 where the second and the third column of each attribute are processed differently and the first one, i.e. the original one is kept.
One can simply inverse the order of the attributes via 'last-first'.
After appyling the filter, the index of the class attribute is the last attribute.

Valid options are:

 -R <index1,index2-index4,...> Specify list of columns to copy. First and last are valid indexes. (default first-last)

@author FracPete (fracpete at waikato dot ac dot nz) @version $Revision: 6249 $

                            int[] shuff_withClass = new int[filteredInstances.numAttributes()];
                            for (int jj = 0; jj < filteredInstances.numAttributes()-1; jj++){shuff_noClass[jj]=String.valueOf(jj);}
                            if (i!=0){Collections.shuffle(Arrays.asList(shuff_noClass), new Random(i));}
                            for (int kk = 0; kk < filteredInstances.numAttributes()-1; kk++) {shuff_withClass[kk]=Integer.valueOf(shuff_noClass[kk]);AttOrder.append((1+shuff_withClass[kk])).append(";");}
                            shuff_withClass[filteredInstances.numAttributes()-1] = filteredInstances.numAttributes()-1; AttOrder.append("last");
                            Reorder reo = new Reorder();
                            reo.setAttributeIndicesArray(shuff_withClass);
                            Instances shuffledInstances;
                            reo.setInputFormat(filteredInstances);
                            shuffledInstances = Filter.useFilter(filteredInstances, reo);
                            validShuff = Filter.useFilter(validationSet, reo);


                            double avgRoc = 0;
                            double avgAcc = 0;

Examples of weka.filters.unsupervised.attribute.Reorder

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