Options for Strengthening the Ensemble of Hypotheses under Uncertainty of the Objective Learning Function Formation

Abstract

Intelligent learning systems traditionally consist of three main components: a student model, which is a block with information about the student; a model of the learning process that sets the form for presenting information to the student and the type of quality assessment of the student’s activity; the model interface as a link between the expert block of the intelligent learning system and other learning algorithms in the components of educational systems. These parts are integral elements in their work on the formation of knowledge bases, learning strategies, assessment procedures, as well as in organizing interaction between the system and users. The paper considers the problem of finding an objective function when setting up a learning system by introducing the possibility of strengthening an ensemble of hypotheses using a learning function, the set of values of which is formed on the basis of the weighted costs of the initial hypotheses, taking into account their own weights and the results of the classification of the corresponding examples.