RUS  ENG
Full version
JOURNALS // Sistemy i Sredstva Informatiki [Systems and Means of Informatics] // Archive

Sistemy i Sredstva Inform., 2015 Volume 25, Issue 3, Pages 60–77 (Mi ssi417)

This article is cited in 1 paper

Building superposition of deep learning neural networks for solving the problem of time series classification

M. S. Popovaa, V. V. Strijovb

a Moscow Institute of Physics and Technology, 9 Institutskiy Per., Dolgoprudny, Moscow Region 141700, Russian Federation
b Dorodnicyn Computing Center, Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 40 Vavilov Str., Moscow 119333, Russian Federation

Abstract: This paper solves the problem of time series classification using deep learning neural networks. The paper proposes to use a multilevel superposition of models belonging to the following classes of neural networks: two-layer neural networks, Boltzmann machines, and autoencoders. Lower levels of superposition extract informative features from noisy data of high dimensionality, while the upper level of superposition solves the problem of classification based on these extracted features. The proposed model was tested on two samples of physical activity time series. The classification results obtained by the proposed model in the computational experiment were compared with the results which were obtained on the same datasets by foreign authors. The study showed the possibility of using deep learning neural networks for solving problems of physical activity time series classification.

Keywords: classification; time series; deep learning neural networks; model superposition; feature extraction.

Received: 25.05.2015

DOI: 10.14357/08696527150304



Bibliographic databases:


© Steklov Math. Inst. of RAS, 2024