Clustering for Probability Density Functions by New -Medoids Method

Table 4

Comparison of all algorithms in example 4 (116 traffic images).

Algorithm	Average adjusted Rand index	Average computational time (seconds)	Average iteration number
Proposed -medoids	1.00	0.394	1
Random -medoids	0.71	0.146	1
Nonhierarchical approach	0.15	0.576	35
SU	0.84	2.485	8