Comparative analysis of stock price ARIMA and LSTM forecasting methods
Articles
Aivaras Bielskis
Vilnius University
https://orcid.org/0000-0002-7626-6951
Igoris Belovas
Vilnius University
https://orcid.org/0000-0002-0478-1102
Published 2022-12-10
https://doi.org/10.15388/LMR.2022.29755
pdf

Keywords

time series
neural networks
forecasting
ARIMA
SARIMA
LSTM

How to Cite

Bielskis, A. and Belovas, I. (2022) “Comparative analysis of stock price ARIMA and LSTM forecasting methods”, Lietuvos matematikos rinkinys, 63(B), pp. 21–27. doi:10.15388/LMR.2022.29755.

Abstract

In the work, relevant methods of stock price forecasting are applied and compared: statistical time series (ARIMA, SARIMA) and neural network-based (LSTM). The results of stock price (Amazon, Apple, Google, Netflix, and Tesla companies) simulations are evaluated using MAE and MRE measures. The conclusions obtained in the work made it possible to identify shortcomings of the approaches and specify guidelines for improvements and further research.

pdf

Downloads

Download data is not yet available.