Figure 3

Validation of the cluster assignments for the acetaminophen data. modk-prototypes clustering of the acetaminophen data was performed 100 times at k = 3 using equal weighting of the microarray, clinical chemistry and histopathology domain data. The necrosis of the centrilobular region of the rat liver histopathology observation was removed from the data prior to clustering and used as an external indicator of cluster assignment validation. The adjusted Rand index (x axis) was computed for each clustering of the data and graphed by count (y axis) of cluster assignments scored with the range of the index.