16.5. Validity of Individual Clusters
There are two cases in which individual cluster validity may be of interest. One is when we want to test whether a given subset of X forms a “good” cluster. “Good” in this case is interpreted in terms of compactness, with respect to its own data, and isolation with respect to the other vectors of X. The other case concerns the validation of a cluster resulting from the application of a clustering algorithm. To this end, both external and internal criteria may be used.
16.5.1. External Criteria
In this section we consider hard clusters and ordinal-type proximity matrices [Bail 82, Jain 88]. The goal is to test whether a given subset of X forms a compact and well-separated cluster. In [Bail 82], two indices ...
Get Pattern Recognition, 4th Edition now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
