Title: Implementation of support vector regression and SARIMA for stock price prediction: Pfizer and AstraZeneca (2017-2025)

Authors: Nazhifa Capella Az'Zahra; Widiarti Putri Kinasih; Syifana Hilmi Muthmainnah; Maghfira Azmi Aqila; Lutfiyah Nur Azizah

Addresses: Faculty of Agricultural Industrial Technology, Department of Agro-Industrial Technology, Universitas Padjadjaran, Jatinangor, West Java, 45363, Indonesia ' Faculty of Agricultural Industrial Technology, Department of Agro-Industrial Technology, Universitas Padjadjaran, Jatinangor, West Java, 45363, Indonesia ' Faculty of Agricultural Industrial Technology, Department of Agro-Industrial Technology, Universitas Padjadjaran, Jatinangor, West Java, 45363, Indonesia ' Faculty of Agricultural Industrial Technology, Department of Agro-Industrial Technology, Universitas Padjadjaran, Jatinangor, West Java, 45363, Indonesia ' Faculty of Agricultural Industrial Technology, Department of Agro-Industrial Technology, Universitas Padjadjaran, Jatinangor, West Java, 45363, Indonesia

Abstract: This study compares the seasonal autoregressive integrated moving average (SARIMA) and support vector regression (SVR) methods for forecasting the stock prices of Pfizer Inc. and AstraZeneca Plc from 2017 to 2025. Monthly closing prices were obtained from Yahoo Finance and analysed in R Studio. SARIMA successfully captured the seasonal behaviour of the data, whereas SVR was applied to model potential non-linear patterns. Model performance was evaluated using mean absolute percentage error (MAPE) and root mean square error (RMSE). The results show that SARIMA achieved higher accuracy, with MAPE values of 4.97% for AstraZeneca and 4.95% for Pfizer, outperforming the SVR model in both MAPE and RMSE metrics. Forecasts generated using the SARIMA (0,1,1)(1,0,0)[12] model indicate a steady increase in AstraZeneca's stock price throughout 2026. In contrast, Pfizer's stock price is projected to remain stable at approximately $31 per share. Overall, the SARIMA model proved to be more effective in this study for predicting pharmaceutical stock prices characterised by seasonal patterns and moderate volatility.

Keywords: stock price forecasting; SARIMA; support vector regression; time series analysis; pharmaceutical sector; MAPE; RMSE; AstraZeneca; Pfizer; financial forecasting.

DOI: 10.1504/IJMABS.2026.155338

International Journal of Markets and Business Systems, 2026 Vol.7 No.2, pp.138 - 162

Received: 13 Oct 2025
Accepted: 30 Nov 2025
Published online: 30 Jul 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article