Cfrgtkky

While implementing Cobweb clustering via weka java API, how to select attributes based on which clustering is to be performed. I have tried following line of code:

public String clustering(Instances data) {

        try {

            Cobweb cw = new Cobweb();

            cw.buildClusterer(data);



            Instance current;

//            while ((current = loader.getNextInstance(structure)) != null) {

//                cw.updateClusterer(current);

//            }



            cw.updateFinished();



            //Crear el objeto para la validacion del modelo con Clustering

            ClusterEvaluation clusterEval = new ClusterEvaluation();



            clusterEval.setClusterer(cw);



            String resCluster = "<br><b><center>Resultados Cluster"

                    + "</center><br>========<br>Modelo generado indica los "

                    + "siguientes resultados:<br>========<br></b>";



            resCluster = resCluster + ("<b>1. Agrupacion:</b> "

                    + clusterEval.getNumClusters() + "<br>");



            return resCluster;



        } catch (Exception e) {

            return "El error es" + e.getMessage();

        }



    }

but an error is generated and it says

weka.clusterers.Cobweb: Cannot handle any class attribute!