Muhamad Fadhil AC, Atiqah Meutia Hilda

Sistem Informasi Prediksi Harga Saham Bank Syariah Menggunakan Metode Arima dan Sarima dengan Antarmuka Visual

Introduction

Sistem informasi prediksi harga saham bank syariah menggunakan metode arima dan sarima dengan antarmuka visual. Sistem informasi prediksi harga saham Bank Syariah (BRIS) berbasis web. Menggunakan metode ARIMA & SARIMA dengan antarmuka visual interaktif untuk mendukung keputusan investasi data-driven.

Abstract

Stock price movements are highly dynamic, requiring prediction approaches that are not only accurate but also easy for users to understand. This study focuses on the development of abased stock price prediction information system for Bank Syariah Indonesia Tbk (BRIS) using a time series forecasting approach. The data used consist of historical BRIS stock prices (open, high, low, close, and volume) obtained from Investing.com and processed through data cleaning, normalization, and preparation to meet time series modeling assumptions. The prediction models applied in this study are ARIMA and SARIMA, with parameter selection based on ACF and PACF analysis. Model performance was evaluated using Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) to determine the accuracy level of the predictions. The evaluation results indicate that the ARIMA model outperforms the SARIMA model, achieving an MAE of  0,1266, RMSE of 0,1519, while the SARIMA model records an MAE of 0,1811, RMSE of 0,1955. The best model was then integrated into a web-based information system using Flask and React.js, which provides visualization of prediction results through interactive charts and comparisons with actual data. The system displays stock price prediction results in the form of interactive charts alongside actual data comparisons, aiming to help users understand stock price trends and support more objective, data-driven investment decisions.

Review

This study presents a timely and relevant contribution to the field of financial forecasting by developing an information system for predicting the stock prices of Bank Syariah Indonesia Tbk (BRIS). The core strength of the paper lies in its dual focus: employing established time series forecasting models (ARIMA and SARIMA) and integrating the best-performing model into a user-friendly, web-based visual interface. The authors effectively highlight the need for prediction approaches that are not only accurate but also easily interpretable for users, a critical aspect often overlooked in purely technical forecasting research. The clear methodology, including data cleaning, normalization, and parameter selection via ACF/PACF analysis, provides a solid foundation for the predictive models. The methodological execution and evaluation are commendably detailed. The study meticulously compares ARIMA and SARIMA models, evaluating their performance using standard metrics such as Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). The reported results, with ARIMA outperforming SARIMA (MAE of 0.1266, RMSE of 0.1519 for ARIMA versus MAE of 0.1811, RMSE of 0.1955 for SARIMA), clearly establish the superior model for integration. The subsequent integration of the optimal ARIMA model into a Flask and React.js based web system is a practical and well-executed step, providing interactive charts for visualizing prediction results against actual data. This practical application significantly enhances the utility of the research, moving beyond theoretical model comparison to a tangible tool that supports data-driven investment decisions. While the study successfully demonstrates the utility of ARIMA for BRIS stock prediction and delivers a functional information system, there are avenues for further exploration. The primary limitation of ARIMA/SARIMA models is their reliance on historical patterns, which may not fully capture sudden market shifts or external influencing factors such as geopolitical events, economic news, or sentiment. Future work could benefit from incorporating exogenous variables or exploring more advanced machine learning and deep learning models (e.g., LSTMs, Transformers) that are adept at handling complex, non-linear dependencies. Additionally, expanding the scope to include a portfolio of Islamic bank stocks or comparing the system's performance against expert human analysis could provide further validation. Nevertheless, this paper provides a robust framework and a valuable tool for stakeholders in the Sharia finance sector, making a significant step towards accessible and data-informed investment strategies.

Full Text

Comments

You need to be logged in to post a comment.