Multi center polyhedral conic classifiers that can classify complex data

Saglamlar H.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, vol.36, no.4, pp.1817-1830, 2021 (SCI-Expanded) identifier identifier identifier

  • Publication Type: Article / Article
  • Volume: 36 Issue: 4
  • Publication Date: 2021
  • Doi Number: 10.17341/gazimmfd.799556
  • Journal Name: JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Art Source, Compendex, TR DİZİN (ULAKBİM)
  • Page Numbers: pp.1817-1830
  • Keywords: Polyhedral conic classifiers, multi center polyhedral conic classifiers, kernel SVM, k-means, classification, FUNCTIONS ALGORITHM, CLASSIFICATION
  • Eskisehir Osmangazi University Affiliated: Yes

Abstract

Polyhedral conic classifiers, compared to support vector machines, stand out for their success while keeping simplicity. In these classifiers, a conic function with a vertex point is used. A carefully chosen vertex point provides compact decision boundaries for positive samples that are clustered together. In cases where positive data belonging to the same class are clustered in different regions, a single classifier is not enough, and more than one classifier is needed. In this study, a novel multicenter polyhedral conic classifier (MCPCC) is developed to use only one classifier to classify positive data clustered in different regions. Experiments are performed using proposed method and related methods for comparison. In the results, it is demonstrated that the new MCPCC method gives successful and promising results.