Markov Chain-Based Stock Forecasting: A Case Study on ADVANC Using Moving Averages

Applying a discrete-time Markov chain to stock prices, which are inherently continuous in nature, necessitates discretization-a critical step that significantly impacts forecasting performance. This research aims to forecast future states of stock prices using a discrete-time Markov chain and employs various types of moving averages: simple moving average (SMA), weighted moving average (WMA), and exponential moving average (EMA). Focusing on ADVANC stock, listed on the Stock Exchange of Thailand, as a case study and using a training dataset, our findings reveal that the 4-day WMA yields the highest accuracy at 65.5%, closely followed by the 2-day EMA at 64.4%. An analysis of all 10 models indicates that both the period and the type of moving average significantly influence forecasting accuracy. However, for ADVANC stock, EMA appears to be more suitable for discretization. Additionally, this study proposes a step-by-step process for stochastic analysis.