BetaSutte: Applying Novelty in Data Forecasting with the Modified Trend-Augmented α-Sutte Indicator, A Case Study on Bank Mandiri (BMRI) Stock Prices

Abstract

This study assesses the effectiveness of the BetaSutte forecasting model, an enhanced version of the α-Sutte Indicator, dubbed the Modified Trend-Augmented α-Sutte Indicator, in forecasting the stock prices of Bank Mandiri (BMRI). Data from investing.com, spanning January 2018 to December 2023, was divided into training and testing subsets to both develop and validate the forecasting model, ensuring it performs well across unseen data. BetaSutte builds on the foundational α-Sutte by integrating advanced trend analysis, mitigating the influence of outliers, and utilizing automatic parameter optimization to boost forecasting precision. The efficacy of BetaSutte is evaluated against well-established models such as SVR, XGBoost, and ARIMA. ARIMA was chosen for its detailed management of time-series data via autoregressive, differencing, and moving average components. In contrast, SVR and XGBoost are recognized for their strong predictive performance. The performance of these models was gauged using Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). These metrics shed light on the extent of forecasting errors and the percentage of relative errors, respectively, providing a comprehensive view of each model’s predictive accuracy for BMRI stock prices. The results demonstrated that BetaSutte outstripped the other models in terms of RMSE and MAPE, highlighting its enhanced ability to accurately reflect the dynamics of BMRI’s stock prices with greater precision and dependability. This establishes BetaSutte as a formidable tool in financial forecasting, particularly valuable in environments characterized by volatile market conditions and unpredictable data patterns.