Analysis of the Exchange Rate on the Thai Baht Against the Chinese Yuan Using a Support Vector Machine and Firefly Algorithm

Abstract

       This research constructed an optimal model to forecast the exchange rate of the Thai Baht against the Chinese Yuan by Support Vector Machine model and firefly algorithm. The software used for the construction was R program. The data used for modelling was secondary data collected from Bank of Thailand and the Ministry of Commerce, which consisted of the exchange rate of the Thai Baht against the Chinese Yuan, policy interest rate (per year), the Thai Baht index, import value, export value and international reserve fund. The collection of data was monthly records starting from January 2009 to June 2019, 126 data sets. The first 120 data sets were used for constructing the model and the last 6 data sets were used to verify the model. It was found that there were 3 factors which affected the exchange rate of the Thai baht against the Chinese yuan, with 5% statistical significance. The same direction factors were policy interest rate and import value and the opposite direction factor was international money fund. The optimal support vector machine model obtained was eps-regression type with radial basis function kernel, which had gamma parameter , epsilon parameter  and cost value parameter . The model verification showed that the obtained model provided the root mean square error 0.1518 only whereas the classical multiple linear regression model provided the root mean square error 0.2614.