<![CDATA[This study aims to analyse the performance of a stock price forecasting model based on Geometric Brownian Motion (GBM) modified with a Markov Switching (MS) approach. The research gap addressed is the limitation of the classical GBM model, which assumes constant volatility and is therefore unable to capture sudden changes in market regimes. To address these limitations, this study proposes a hybrid GBM-MS model as its main scientific contribution, in whichthe drift and volatility parameters are dynamically estimated following changes in market conditions through a switching mechanism between regimes. Parameter estimation is performed using the Hidden Markov Model. The model's performance is compared with the classical GBM as a benchmark. The research uses daily closing price data of PT Bank Central Asia Tbk. (BBCA) shares for the period 1 July – 30 December 2024. The results show that the hybrid GBM-MS model provides better forcasting accuracy with a MAPE value of 2.25% on training data and 1.38% on testing data, lower than the classical GBM model. These findings confirm that the integration of Markov Switching enhances the model's adaptability in capturing structural changes and market volatility. Practically, the hybrid GBM-MS model can be used as a more reliable forecasting and risk management tool to support investment decision-making, especially in dynamic and unstable market environments.]]>
Copyrights © 2025
