i've run k-means cluster analysis in r , have identified 6 unique clusters. when assign cluster number raw data, see there overlaps of variables clusters. did expect because have ~20 variables.
what can manually @ frequency of each variable per cluster , compare population (the other clusters) see if there bias particular variable.
is there anyway automate process , show of variables per cluster have bias. allow me compare , contrast clusters determine unique characteristics.
Comments
Post a Comment