Algorithms for constructing of decision trees

Abstract

Decision trees are widely used in different applications for problem solving and for knowledge representation. In the paper algorithms for decision tree constructing with bounds on complexity and precision are considered. In these algorithms different measures for time complexity of decision trees and different measures for uncertainty of decision tables are used. New results about precision of polynomial approximate algorithms for covering problem solving [1, 2] show that some of considered algorithms for decision tree constructing are, apparently, close to unimprovable.