Classification with Incomplete Probabilistic Labeling Based on Manifold Regularization and Fuzzy Clustering Ensemble | Sciact - Система учета научной деятельности ИМ СО РАН

Classification with Incomplete Probabilistic Labeling Based on Manifold Regularization and Fuzzy Clustering Ensemble Научная публикация

Журнал

Pattern Recognition and Image Analysis
ISSN: 1054-6618 , E-ISSN: 1555-6212

Вых. Данные

Год: 2022, Том: 32, Номер: 3, Страницы: 515-518 Страниц : 4 DOI: 10.1134/S1054661822030075

Ключевые слова

cluster ensemble; fuzzy partitioning; low-rank matrix approximation; manifold regularization; probabilistic labeling; weakly supervised learning

Авторы

Berikov V.B. ¹ , Vikent'ev A.A. ¹

Организации

1	Sobolev Institute of Mathematics, Siberian Branch, Russian Academy of Sciences, Novosibirsk, 630090 Russia

Информация о финансировании (1)

Российский научный фонд

22-21-00261

The paper proposes a weakly supervised binary classification method which combines manifold regularization and fuzzy clustering ensemble methodologies. We assume that the class labels can be fully supervised, defined in terms of a probability distribution or not specified at all. The co-association matrix of fuzzy clustering ensemble is used as the similarity matrix. This matrix is represented in a low-rank form that significantly speeds up calculations and saves memory. Numerical experiments using Monte Carlo modeling demonstrate the efficiency of the method.

Berikov V.B. , Vikent'ev A.A.
Classification with Incomplete Probabilistic Labeling Based on Manifold Regularization and Fuzzy Clustering Ensemble
Pattern Recognition and Image Analysis. 2022. V.32. N3. P.515-518. DOI: 10.1134/S1054661822030075 WOS Scopus РИНЦ OpenAlex

Поступила в редакцию:	31 мая 2022 г.
Принята к публикации:	31 мая 2022 г.
Опубликована в печати:	19 окт. 2022 г.
Опубликована online:	19 окт. 2022 г.

Web of science:	WOS:000869886400011
Scopus:	2-s2.0-85140117752
РИНЦ:	49604113
OpenAlex:	W4306745292

БД	Цитирований
РИНЦ	1