Predicting time series data using Hybrid SARIMAX-LSTM algorithm†
Nanyoung Hong1 · Younjae Lee2 · Taewook Lee3
123Department of Statistics, Hankuk University of Foreign Studies
Correspondence to: † This work was supported by Hankuk University of Foreign Studies Research Fund of 2023.
1 Graduate student, Department of Statistics, Hankuk University of Foreign Studies, Gyeonggi-do 17035, Korea.
2 Ph.D. student, Department of Statistics, Hankuk University of Foreign Studies, Gyeonggi-do 17035, Korea.
3 Corresponding author: Professor, Department of Statistics, Hankuk University of Foreign Studies, Gyeonggi-do 17035, Korea. E-mail: twlee@hufs.ac.kr
1 Graduate student, Department of Statistics, Hankuk University of Foreign Studies, Gyeonggi-do 17035, Korea.
2 Ph.D. student, Department of Statistics, Hankuk University of Foreign Studies, Gyeonggi-do 17035, Korea.
3 Corresponding author: Professor, Department of Statistics, Hankuk University of Foreign Studies, Gyeonggi-do 17035, Korea. E-mail: twlee@hufs.ac.kr
Received August 2, 2023; Revised August 21, 2023; Accepted August 28, 2023.
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
- The seasonal autoregressive integrated moving average (SARIMA) model has long been used as a method for predicting time series data. As LSTM algorithm, which is an artificial neural network suitable for time series prediction, was devised, the prediction performance of SARIMA model and LSTM algorithm was often compared. In this paper, we propose a Hybrid SARIMAX-LSTM algorithm that combines SARIMAX, a traditional time series model that considers seasonality and exogenous variables, with LSTM algorithm. Through the results of simulation and empirical studies, we intend to prove that Hybrid SARIMAX-LSTM algorithm is effective in improving the prediction performance and shortening the analysis time compared to the existing LSTM algorithm.
- Keywords : Exogenous variable, LSTM, SARIMA, seasonality, time series
Full Text(PDF) Free
