Modelling The US Dollar Index Using Continuous Hidden Markov

David Vijanarco Martal (School of Data Science, Mathematics and Informatics, IPB University)
Berlian Setiawaty (School of Data Science, Mathematics and Informatics, IPB University)
Retno Budiarti (School of Data Science, Mathematics and Informatics, IPB University)

Journal of the Indonesian Mathematical Society
Vol. 32 No. 3 (2026): SEPTEMBER

This study employs the continuous hidden Markov model (HMM) to model the index data of the US Dollar index from 2018 to 2024. HMM is used to predict and analyze hidden patterns that generated the data. The data is modelled using a continuous HMM with 11 hidden states and lognormal distributions with different parameters for each hidden state. The accuracy of the continuous HMM is measured by MAPE. The MAPE value for both training and testing data is very low, less than 4%. This means that the continuous HMM can be used to model the data accurately. The plot shows accurate predictions between simulated data and real data, and furthermore, the model can capture the fluctuations of the data.