Stock Price Forecasting Model Using Short Cross Association of Logical Fuzzy Relations in Fuzzy Time Series

Abstract

Forecasting stock prices plays a crucial role in financial engineering and data-driven decision-making systems in the industrial world. The accuracy of stock price predictions is essential for investors, financial analysts, and infrastructure project managers to mitigate investment risks and optimize business strategies. Fuzzy Time Series (FTS) has become one of the popular methods for time series forecasting due to its ability to handle uncertainty and nonlinear patterns that often arise in financial and engineering systems. However, conventional FTS methods still face challenges in forming Fuzzy Logical Relationships (FLR), which can affect the accuracy of prediction results. This study proposes the Short Cross-Association Fuzzy Logical Relationship (SCA-FLR) approach to improve forecasting accuracy by considering influential factors in FLR formation. This method is applied to stock price data of PT Wijaya Karya (Persero) Tbk (WIKA.JK), using the stock closing price as the main factor and the highest stock price as the influencing factor. The forecasting results show an Average Forecasting Error Rate (AFER) of 2.97%, indicating excellent prediction accuracy. The findings of this study contribute to the development of forecasting systems in financial engineering, risk management, and industrial decision-making optimization. The application of this method can be extended to various engineering fields involving time series analysis.