Work

An Interpretable Belief Rule Base with Dynamic Reference Value for Stock Price Trend Forecasting

Year: 2026

Type: article

Abstract: Accurately predicting stock price trends is crucial for the decision-making process in the field of financial intelligence.However, the non-stationarity and high volatility of financial data make its prediction a key challenge.The existing models generally struggle to balance prediction accuracy and model interpretability.Therefore, this study proposes an interpretable belief rule base with dynamic reference value (I-BRB-DR) framework for intelligent stock price trend forecasting.Firstly, a reference value dynamic adjustment mechanism is designed to ensure adaptability to sudden market changes by analyzing real-time market trends and autonomously updating the threshold of the belief rule base.Secondly, a set of interpretability constraint strategies are introduced in the parameter optimization procedure to ensure logical consistency and model interpretability of the model before and after optimization.Finally, the efficacy of I-BRB-DR is validated using historical stock market data in a case study.The results indicate that the model can not only accurately capture complex market patterns, but also provide clear rule-based explanations for its forecasts.This study proposes an efficient, interpretable, and adaptable forecast model, which provides a powerful tool for the field of financial intelligence and has the potential to expand to a wider range of time series prediction tasks.

Source: KSII Transactions on Internet and Information Systems

DOI: 10.3837/tiis.2026.01.006