New internal clustering validation measure for contiguous arbitrary?shape clusters

In this study a new internal clustering validation index is proposed. It based on measure of the uniformity data in clusters. uses local density each cluster, particular, normalized variability within clusters to find ideal partition. The validity allows it capture spatial pattern and obtain right number an automatic way. This approach, unlike traditional one that usually identifies well-separated compact clouds, works with arbitrary-shape may be contiguous or even overlapped. proposed has been evaluated nine artificial sets, different cluster distributions increasing classes, three highly nonlinear 17 real sets. compared well-known indices very satisfactory results. proves including definition useful identify partition different-size