FEATURES OF RANDOM LOGICAL CLASSIFICATION TREES IN PATTERN RECOGNITION PROBLEMS
Igor Povkhan · Scientific notes of Taurida National V I Vernadsky University Series Technical Sciences · 2019
The work raises an important question of pattern recognition theory -the use of methods and algorithms for constructing random logical classification trees.The principal features of random classification trees are considered, i.e. logical trees in which the selection (generation) of vertices at an arbitrary stage of tree construction occurs randomly.The algorithm proposed in this paper allows the construction of a random logical tree, and generate whole sets of logical trees of different structures (complexity), among which you can choose the most optimal for this problem.Emphasizes the importance of the issue in the application of random Boolean trees for the solution of pattern recognition problems -the question of selection (and possible adjustment) is the most efficient tree among the many built of random logical trees.The work is recorded that random logical tree has its significant advantages (software, the ease of construction of a classification tree, reducing the time of the generation of the logical tree, the ability to evaluate and select the most appropriate classification tree from the set was built) and significant drawbacks (not the optimal structure, the hardware cost of generation is not the optimal classification tree, guaranteed by the complicated structure and large information capacity, the need for additional evaluation phase and selection).A simple, efficient, cost-effective method of constructing a random logical classification tree training sample allows you to provide the necessary speed, the level of complexity of the recognition scheme, which guarantees a simple and complete recognition of discrete objects.The paper deals with tree-like recognition schemes.