QUESTION OF THE OPTIMALITY CRITERION OF A REGULAR LOGICAL TREE BASED ON THE CONCEPT OF SIMILARITY

Abstract

This work is devoted to the problem of logical trees – the question of the optimality criterion of a regular logical tree based on the concept of similarity. Solving this issue will allow us to develop effective methods and algorithms for minimizing the structure of logical classification trees. The minimal structure of the logical classification tree provides a fast and efficient mechanism for classifying discrete objects. The obtained result is fundamentally important in the problem of evaluating the effect of permutation of the tiers of the maximum logical classification tree and the question of the structural complexity of the obtained logical trees. Fixing the initial training sample in the form of a logical tree, creates a fixed tree-like data structure, which to some extent provides even compression and transformation of the initial data of the training sample, and therefore allows significant optimization and saving of hardware resources of the information system. The work is relevant for all image recognition approaches in which the resulting classification scheme can be represented as a logical tree.

Key words: tasks of pattern recognition, logical tree, graph-scheme of the model.