MODEL GEOMETRIC BROWNIAN MOTION TERMODIFIKASI KALMAN FILTER UNTUK PREDIKSI SAHAM KONSUMEN DAN IMPLIKASINYA TERHADAP STRATEGI INVESTASI KELUARGA DI INDONESIA

Abstract

Stock investment is increasingly in demand by investors because of its high profit potential, but predicting stock price movements is still difficult due to its volatile nature. The Geometric Brownian Motion Model (GBM) is commonly used for this purpose, as it captures the stochastic dynamics of stock prices, assuming a normal log-return distribution and constant volatility. However, prediction accuracy decreases over time due to the dynamic nature of the stock market. To improve prediction accuracy, the Kalman Filter is used to iteratively adjust parameters in the GBM model, resulting in a more flexible and accurate forecasting approach. Research by Maulidya et al. revealed that the combination of GBM and Kalman Filter produced a low Mean Absolute Percentage Error (MAPE) of 0.0674%, indicating high prediction accuracy. This study aims to develop a stock price prediction model for PT Unilever Indonesia Tbk (UNVR) using GBM modified with the Kalman Filter, resulting in a model that is more representative and adaptive to market changes. The methodology includes data collection, return calculation, parameter estimation, and model construction, with the results showing a MAPE of 6.8%, outperforming the traditional GBM model. The study concludes that the modified GBM-KF model is effective for short-term prediction, highlighting its ability to adapt to market fluctuations despite challenges posed by non-linearity and extreme market conditions.