Implementation of Gaussian hidden Markov model (GHMM) to predict gold prices in Indonesia

The Gaussian Hidden Markov Model (GHMM) is a Hidden Markov Model (HMM) that uses the Gaussian Distribution or Normal Distribution as the probability of its observations. GHMM is an effective method used in various fields. In this study, GHMM is used to predict gold prices in Indonesia, which is the best asset for investment. Gold price prediction is done in two stages. The first stage is to determine the optimum number of states, where the smallest AIC and BIC values are selected in the training data and evaluate the MAPE values on the data testing. The second stage is to predict the price of gold in Indonesia for ten days by calibrating the parameters using the Baum-Welch Algorithm. The results of the study show that the optimum number of states from the training data based on the smallest AIC and BIC values is 5 (five). Then the evaluation of the model against data testing obtained a MAPE value of 0,55%. Then the prediction results obtained have a MAPE values of 0,30% or less than 10%. Therefore, it is concluded that GHMM can predict gold prices in Indonesia well.