search for




 

Time series forecasting using hybrid wavelet-SARIMA model
Journal of the Korean Data & Information Science Society 2022;33:829-43
Published online September 30, 2022;  https://doi.org/10.7465/jkdi.2022.33.5.829
© 2022 Korean Data and Information Science Society.

Byeongjin Yoon1 · Bogeun Sim2 · Haejune Oh3

123Department of Information and Statistics, Gyeong National University
Correspondence to: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NRF-2020R1G1A1A01014362).
1 Graduate student, Department of Information and Statistics, Gyeongsang National University, Jinju 52828, Korea.
2 Graduate student, Department of Information and Statistics, Gyeongsang National University, Jinju 52828, Korea.
3 Assistant professor, Department of Information and Statistics, Gyeongsang National University, Jinju 52828, Korea. E-mail: haejune.oh@gmail.com
Received August 16, 2022; Revised September 5, 2022; Accepted September 13, 2022.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
In this study, we propose a hybrid model for analysis and prediction of time series data based on the wavelet transform that can effectively analyze and predict complexity such as non-stationary and non-linearity of time series data. In order to improve prediction performance, the coefficients generated by discrete wavelet decomposition are denoised by threshold processing with the Garrote threshold function, and the threshold-treated coefficients are predicted with the SARIMA model to obtain a prediction value by combining each coefficient. We predict the weekly averaged temperature in Jinju-si using this hybrid model. And the prediction performance of the proposed hybrid model is compared with those of SARIMA model, wavelet denoising-SARIMA model and wavelet decomposition-SARIMA model.
Keywords : Denoising, discrete wavelet transform, SARIMA model, time series forecasting, wavelet decomposition.