Up to date, research on automatic classification focused mainly on the
efficiency of algorithms in regard to a given aspect of a dataset. The
issue of classification of a given dataset in regard to various aspects is
generally not addressed. In this chapter, a multi-facet hierarchical classification
technique based on the single-link clustering is proposed. First,
the single-link clustering is formally presented. Then, the concepts underlying
the generic hierarchical classification technique are given. Next,
analysis domains modeling a given facet of a dataset are described. A
new language devoted to generate analysis domains is presented. Further,
classification of analysis domains is discussed. Finally, examples
of applications of the generic hierarchical classification are given.
|