Involment of additional constraints of subject domain in machine learning problems

The study deals with the concideration of a range of issues related to the involving additional knowledge about subject domain when solving machine learning problems. The techniques for accounting such knowlwdge based on the modification of classical methods of classification, clustering, associative rule discovery are described. It is concluded that the analysis of bachgroung knowledge is able to increase reliability and accuracy of classical machine learning methods, although their modification, taking into account additional constraints, sometimes turns out to be a rather time-consuming procedure. In addition, different types of machine learning problems require different types of additional constraints to increase the reliability and accuracy of their results. This provides certaint difficulties in solving complex problems that require gradual involvement of various types of additional constraints.