PREDIKSI ARAH HARGA EMAS DENGAN RANTAI MARKOV WAKTU DISKRET DUA STATE

Abstract

Gold plays many important roles, including serving as a store of value, a reserve asset for nations, an economic indicator, and an easily accessible investment instrument. These functions make gold price forecasting a valuable tool for investors, policymakers, and central banks in making decisions regarding the purchase and sale of gold. This study employs a two-state discrete-time Markov chain model to predict the direction of gold price trends, providing information that can support short-term (daily) buy-and-sell decisions. The results indicate that, in the long run, the probabilities of gold prices increasing and decreasing reach their steady-state values, while the average first-passage time between states is approximately two periods. Compared with machine learning models, which are often considered black-box approaches, the Markov chain model is easier to understand and offers a transparent modeling process