KFU. Card publication. Reduction of the dimension of neural network models in problems of pattern recognition and forecasting

KAZAN
FEDERAL UNIVERSITY

- RUS
- 中文
- ES

REDUCTION OF THE DIMENSION OF NEURAL NETWORK MODELS IN PROBLEMS OF PATTERN RECOGNITION AND FORECASTING

Form of presentation

Articles in international journals and collections

Year of publication

2017

��

��

Authors, employees of KFU

Bochkarev Vladimir Vladimirovich, author

Bibliographic description in the original language

Nasertdinova A.D, Bochkarev V.V., Reduction of the dimension of neural network models in problems of pattern recognition and forecasting//Journal of Physics: Conference Series. - 2017. - Vol.929, Is.1. - Art. � 012038.

Annotation

Deep neural networks with a large number of parameters are a powerful tool for solving problems of pattern recognition, prediction and classification. Nevertheless, overfitting remains a serious problem in the use of such networks. A method of solving the problem of overfitting is proposed in this article. This method is based on reducing the number of independent parameters of a neural network model using the principal component analysis, and can be implemented using existing libraries of neural computing. The algorithm was tested on the problem of recognition of handwritten symbols from the MNIST database, as well as on the task of predicting time series (rows of the average monthly number of sunspots and series of the Lorentz system were used). It is shown that the application of the principal component analysis enables reducing the number of parameters of the neural network model when the results are good.

Keywords

neural networks, principal component analysis, reduction of the dimension, MNIST

The name of the journal

Journal of Physics: Conference Series

URL

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85039072328&doi=10.1088%2f1742-6596%2f929%2f1%2f012038&partnerID=40&md5=c91b8b82d293cd298f12f06a5878020a

Please use this ID to quote from or refer to the card

https://repository.kpfu.ru/eng/?p_id=174165&p_lang=2

Full metadata record

Field DC	Value	Language
dc.contributor.author	Bochkarev Vladimir Vladimirovich	ru_RU
dc.date.accessioned	2017-01-01T00:00:00Z	ru_RU
dc.date.available	2017-01-01T00:00:00Z	ru_RU
dc.date.issued	2017	ru_RU
dc.identifier.citation	Nasertdinova A.D, Bochkarev V.V., Reduction of the dimension of neural network models in problems of pattern recognition and forecasting//Journal of Physics: Conference Series. - 2017. - Vol.929, Is.1. - Art. � 012038.	ru_RU
dc.identifier.uri	https://repository.kpfu.ru/eng/?p_id=174165&p_lang=2	ru_RU
dc.description.abstract	Journal of Physics: Conference Series	ru_RU
dc.description.abstract	Deep neural networks with a large number of parameters are a powerful tool for solving problems of pattern recognition, prediction and classification. Nevertheless, overfitting remains a serious problem in the use of such networks. A method of solving the problem of overfitting is proposed in this article. This method is based on reducing the number of independent parameters of a neural network model using the principal component analysis, and can be implemented using existing libraries of neural computing. The algorithm was tested on the problem of recognition of handwritten symbols from the MNIST database, as well as on the task of predicting time series (rows of the average monthly number of sunspots and series of the Lorentz system were used). It is shown that the application of the principal component analysis enables reducing the number of parameters of the neural network model when the results are good.	ru_RU
dc.language.iso	ru	ru_RU
dc.subject	neural networks	ru_RU
dc.subject	principal component analysis	ru_RU
dc.subject	reduction of the dimension	ru_RU
dc.subject	MNIST	ru_RU
dc.title	Reduction of the dimension of neural network models in problems of pattern recognition and forecasting	ru_RU
dc.type	Articles in international journals and collections	ru_RU