Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций
Конференция: Siberian Scientific Workshop on Data Analysis Technologies with Applications, SibDATA 2020
Год издания: 2020
Ключевые слова: decision trees, gradient boosting, process disruptions, random forest
Аннотация: This paper considers the task of elaborating the tools that enable early detection of process disruptions in aluminum production using the technology of decision trees. The suggested method to forecast the process disruptions are based on the data on daily average process indicators. The method includes a necessary stage of preliminary processing of inputs and consequent construction of a math model. The study defined the most informative properties, solved the problem of unbalanced data, and compared approaches based on decision trees. The quality metrics revealed the most effective method to solve the set task. Copyright © 2020 for this paper by its authors.
Издание
Журнал: CEUR Workshop Proceedings
Выпуск журнала: Vol. 2727
Номера страниц: 75-82
ISSN журнала: 16130073
Издатель: CEUR-WS
Персоны
- Lugovaya N. (Siberian Federal University, 26, Kirenskogo str., Krasnoyarsk, 660074, Russian Federation)
- Mikhalev A. (Siberian Federal University, 26, Kirenskogo str., Krasnoyarsk, 660074, Russian Federation)
- Penkova T. (Institute of Computational Modelling of the Siberian Branch, Russian Academy of Sciences, 50/44 Akademgorodok, Krasnoyarsk, 660036, Russian Federation)
Вхождение в базы данных
Информация о публикациях загружается с сайта службы поддержки публикационной активности СФУ. Сообщите, если заметили неточности.