UDC 519.7

Algorithmic aspects of determining the depth functions in selecting the optimal hypothesis for data classification problems

/ O.A. Galkin // Kibernetika i sistemnyi analiz. — 2016. — Vol. 52, N 5. — P. 43–55.

The paper analyzes optimal hypothesis selection in classification problems based on the hypothesis class distributed with respect to the posterior probability. A method is proposed that is based on the concept of a relative weighted average value and depth functions operating in the space of classification functions. Algorithms are constructed to approximate the relative depth of the data and relative weighted average value providing polynomial approximation to the half-space analogs. Figs: 0. Tabl.: 0. Refs: 10 titles.

Keywords:

weighted average value, relative depth function, optimal Bayesian hypothesis


FULL TEXT

Author(s):

Галкин Александр Анатольевич, кандидат физ.-мат. наук, ассистент кафедры Киевского национального университета имени Тараса Шевченко,
e-mail: galkin.o.a@gmail.com