A new cluster validity measure for simultaneously dealing with datasets having different densities, shapes and sizes and providing optimal partitions

To date, there have been many validity measures proposed for evaluating clustering results in data mining literatures. Most of these popular validity measures do not work well for clusters with different densities, shapes or sizes. They usually have a tendency of ignoring clusters with low densities, small size and irregular shapes. In this paper, we propose a new validity measure AP (Advanced Partition) measure that can address this issue and will also be able to determine the best partitions in the datasets. Our experimental results with both the synthetic and real datasets demonstrate the effectiveness of the proposed validity measure.