Construction of a correct taxonomy algorithm in expansions of the a class of recognition algorithms

✍ Scribed by A.F. Mukhamedgaliev

• Publisher: Elsevier Science
• Year: 1983
• Weight: 288 KB
• Volume: 23
• Category: Article
• ISSN: 0041-5553
• DOI: 10.1016/s0041-5553(83)80022-6

✦ Synopsis

The completeness of a f&mily of taxonomy algorithms, represented in the form of two interconnected subfamilies, is considered.

While constructing the algorithms, defining the completeness of each subfamily, it is snown that a correct algorithm exists in the class of recognition algorithms of the estimate-calculation type.

By a taxonomy problem, we usually, mean the problem of constructing the best (in some sense) splitting of a set of objects into classes.

In the present paper we consider the construction of such a family of taxonomy algorithms whose application to any set of objects gives any a priori defined splitting of the family into non-intersecting classes.

A set of taxonomy algorithms (A,