Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: 11th Computer Science On-line Conference, CSOC 2022

Год издания: 2022

Идентификатор DOI: 10.1007/978-3-031-09073-8_37

Ключевые слова: algorithmic procedure, classification, feature selection, informativity, outliers

Аннотация: When solving classification problems using inductive learning methods, problems may arise in the initial data sample, which lead to the building of patterns with low informativity. First of all, these include the presence of a large number of features that describe observations, as well as the presence of outliers in the training sПоказать полностьюample. To solve the problems under consideration, specially developed algorithmic procedures are proposed that are aimed at preparing the initial data sample for the process of extracting patterns from it with high informativity. A classifier with a high generalizing ability is formed only from highly informative patterns, i.e. high quality classification of new observations. In addition, the paper proposes and tests a heuristic approach for searching for a truncated set of features. Experimental studies are given on a real problem, allowing to determine the feasibility of the proposed algorithmic procedures, which allow solving these problems in the original sample. #CSOC1120. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Журнал: Lecture Notes in Networks and Systems

Выпуск журнала: Vol. 503 LNNS

Номера страниц: 426-434

ISSN журнала: 23673370

Издатель: Springer Science and Business Media Deutschland GmbH

• Kuzmich R.I. (Siberian Federal University, Krasnoyarsk, Russian Federation, Sirius University of Science and Technology, Sochi, Russian Federation)
• Stupina A.A. (Siberian Federal University, Krasnoyarsk, Russian Federation, Reshetnev Siberian State University of Science and Technology, Krasnoyarsk, Russian Federation, Krasnoyarsk State Agrarian University, Krasnoyarsk, Russian Federation)
• Zavalov A.A. (Siberian Federal University, Krasnoyarsk, Russian Federation, Sirius University of Science and Technology, Sochi, Russian Federation)
• Dresvianskii E.S. (Siberian Federal University, Krasnoyarsk, Russian Federation, Sirius University of Science and Technology, Sochi, Russian Federation, University of Cadiz, UCASE Software Engineering Group, University of Cadiz Street, 10, Puerto Real, 11519, Spain)
• Pokushko M.V. (Siberian Federal University, Krasnoyarsk, Russian Federation, Sirius University of Science and Technology, Sochi, Russian Federation, University of Cadiz, UCASE Software Engineering Group, University of Cadiz Street, 10, Puerto Real, 11519, Spain)

• Scopus