Figure 2

Determination of k clusters in the heart disease and acetaminophen data sets using modk-prototypes. The a) heart disease data and b) acetaminophen data were clustered using the modk-prototypes algorithm at values of k increasing from 2 to the N number of samples in the data. DVI_CU (on the y axis) was computed and plotted for the clustering of the data at each value of k (on the x axis). Only k = 2 to 10 is shown.