Assessing Cluster Quality

Clustering results are hard to evaluate, especially for high-dimensional data
and without a priori knowledge of the objects’ distribution, which is quite
common in practical cases. However, assessing the quality of the resulting
clusters is as important as generating the clusters. Given the same data set,
different clustering algorithms with various parameters or initial conditions
will give very different clusters. It is essential to know whether the resulting
clusters are valid and how to compare the quality of the clustering results, so
that the right clustering algorithm can be chosen and the best clustering results
can be used for further analysis.