Group approach to solving the tasks of recognition

Group approach to solving the tasks of recognition Full article

Journal

Yugoslav Journal of Operations Research
ISSN: 0354-0243 , E-ISSN: 2334-6043

Output data

Year: 2019, Volume: 29, Number: 2, Pages: 177-192 Pages count : 16 DOI: 10.2298/YJOR180822032Y

Tags

Classification; Hyper Spectral Image; Recognition; Semi-Supervised Learning

Authors

Amirgaliyev Y. ¹ , Berikov V. ^2,5 , Cherikbayeva L.S. ³ , Latuta K. ¹ , Bekturgan K. ⁴

Affiliations

1	Institute of Information and Computational Technologies, Suleyman Demirel University, Almaty, Kazakhstan
2	Sobolev Institute of Mathematics, SB RAS
3	Alfarabi Kazakh National University, Almaty, Kazakhstan
4	Institute of Automation and Information Technology, Kyrgyzstan
5	Novosibirsk State University, Novosibirsk, Russian Federation

In this work, we develop CASVM and CANN algorithms for semi-supervised classification problem. The algorithms are based on a combination of ensemble clustering and kernel methods. A probabilistic model of classification with the use of cluster ensemble is proposed. Within the model, error probability of CANN is studied. Assumptions that make probability of error converge to zero are formulated. The proposed algorithms are experimentally tested on a hyperspectral image. It is shown that CASVM and CANN are more noise resistant than standard SVM and kNN. © 2019 Faculty of Organizational Sciences, Belgrade. All rights reserved.

Amirgaliyev Y. , Berikov V. , Cherikbayeva L.S. , Latuta K. , Bekturgan K.
Group approach to solving the tasks of recognition
Yugoslav Journal of Operations Research. 2019. V.29. N2. P.177-192. DOI: 10.2298/YJOR180822032Y Scopus OpenAlex

Scopus:	2-s2.0-85068467475
OpenAlex:	W2902851390

DB	Citing
Scopus	2
OpenAlex	2