Icon

Kprototypes

You need to install the clustMixTypepackage in R Node 1Kprototypestransfer the "optimal" number of clusterval$k_optto make the clusterAttachclusterJust showing youthat there is anothercluster algo that canhandle mix dataAssign the clusters 5% of the recordsjust doing thisto time executionto obtain val$k_optbut in real lifeyou should try with all data or at least a representative sampleKprototypes2 is the optimal numberof clusters under silhouette coef the higher the better0.3919564185738614but the value provide littleevidence of good clustersMin Maxyou can try withthe other methodsval$k_optCheck the "optimal"number of clustersbetween k = 2:10 clustersaccording to the silhouette coefval$k_optto variableFile Reader R Snippet Joiner EM (3.7) Weka ClusterAssigner (3.7) Row Sampling R Snippet Normalizer R Snippet Table Rowto Variable You need to install the clustMixTypepackage in R Node 1Kprototypestransfer the "optimal" number of clusterval$k_optto make the clusterAttachclusterJust showing youthat there is anothercluster algo that canhandle mix dataAssign the clusters5% of the recordsjust doing thisto time executionto obtain val$k_optbut in real lifeyou should try with all data or at least a representative sampleKprototypes2 is the optimal numberof clusters under silhouette coef the higher the better0.3919564185738614but the value provide littleevidence of good clustersMin Maxyou can try withthe other methodsval$k_optCheck the "optimal"number of clustersbetween k = 2:10 clustersaccording to the silhouette coefval$k_optto variableFile Reader R Snippet Joiner EM (3.7) Weka ClusterAssigner (3.7) Row Sampling R Snippet Normalizer R Snippet Table Rowto Variable

Nodes

Extensions

Links