83289

Автор(ы): 

Автор(ов): 

4

Параметры публикации

Тип публикации: 

Глава в книге

Название: 

Convex Hulls in Solving Multiclass Pattern Recognition Problem

ISBN/ISSN: 

978-3-030-53551-3

DOI: 

10.1007/978-3-030-53552-0_35

Наименование источника: 

  • Lecture Notes in Computer Science, LNCS 12096

Город: 

  • Cham

Издательство: 

  • Springer Nature Switzerland AG

Год издания: 

2020

Страницы: 

390-401
Аннотация
The paper proposes an approach to solving multiclass pattern recognition problem in a geometric formulation based on convex hulls and convex separable sets (CS-sets). The advantage of the proposed method is the uniqueness of the resulting solution and the uniqueness of assigning each point of the source space to one of the classes. The approach also allows you to uniqelly filter the sourse data for the outliers in the data. Computational experiments using the developed approach were carried out using academic examples and test data from public libraries.

Библиографическая ссылка: 

Гайнанов Д.Н., Рассказова В.А., Чернавин Н.П., Чернавин П.Ф. Convex Hulls in Solving Multiclass Pattern Recognition Problem / Lecture Notes in Computer Science, LNCS 12096. Cham: Springer Nature Switzerland AG, 2020. С. 390-401.