The problems of estimation of the minimum average time complexity of decision trees and design of efficient algorithms are complex in general case. The upper bounds described in Chap. 2.4.3 can not be applied directly due to large computational complexity of the parameter M(z). Under reasonable assumptions about the relation of P and NP, there are no polynomial time algorithms with good approximation ratio [12, 32]. One of the possible solutions is to consider particular classes of problems and improve the existing results using characteristics of the considered classes. © Springer-Verlag Berlin Heidelberg 2011.
ASJC Scopus subject areas
- Library and Information Sciences
- Computer Science(all)
- Information Systems and Management