Abstract: In one embodiment, the present invention includes a method of forming windows corresponding to a data point of a data set, successively expanding the windows, determining a local hill for the windows, re-centering the windows on the local hill, and merging any of the windows within a selected distance of each other. The windows formed may be substantially the same size as a single data point, in one embodiment. The merged windows may be recorded as possible merge points of a hierarchical cluster formed from the data set. Other embodiments are described and claimed.

Abstract: A non-binary affinity measure between any two data points for a supervised classifier may be determined. For example, affinity measures may be determined for tree, kernel-based, nearest neighbor-based and neural network supervised classifiers. By providing non-binary affinity measures using supervised classifiers, more information may be provided for clustering, analyzing and, particularly, for visualizing the results of data mining.