This article is cited in 1 paper

Intellectual Control Systems, Data Analysis

Entropy-based evaluation in classification problems

Yu. A. Dubnov<sup>abc

a</sup> Moscow Institute of Physics and Technology
^b National Research University "Higher School of Economics", Moscow
^c Institute for Systems Analysis of Russian Academy of Sciences

Abstract: The problem of binary classification is considered, an algorithm for its solution based on maximum entropy estimation of the parameters of randomized data models is offered: a detailed description of the model and the entropy estimation method is given, advantages and disadvantages of this approach are described, the results of numerical experiments and comparisons with the classical support vector machine algorithm on the accuracy of classification and depending on the volume of the training sample.

Keywords: machine learning, classification, maximum entropy method.

Presented by the member of Editorial Board: B. M. Miller

Received: 19.07.2018
Revised: 06.11.2018
Accepted: 08.11.2018

DOI: 10.1134/S0005231019030097

Bibliographic databases:

•