Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:D2L-C
Session:
Number:D2L-C-02
Time Series Classification by Neural Network Using Features of Attractors After Smoothing Process
Ryosuke Shimizu , Yoko Uwate , Yoshifumi Nishio,
pp.638-641
Publication Date:12/12/2022
Online ISSN:2188-5079
DOI:10.34385/proc.71.D2L-C-02
PDF download (2.1MB)
Summary:
Time series classification is an important and challenging problem in data analysis. Recently, time series analysis using neural networks~(NN) has attracted much attention. However, the analysis of time series data with complex oscillations is difficult. Therefore, it is important to search for effective features of the data. In this study, we transform the dimensionality of the data and search for features suitable for NN classification. In this study, we investigate the effect of smoothed data on the classification accuracy of the dimensionality reduction method and the features of the data.