การวิเคราะห์อัตราแลกเปลี่ยนเงินบาทด้วยแบบจำลองโครงข่ายประสาทเทียม

Abstract

This research aims to study the variables that affect the exchange rate of the Thai baht against the US dollar using the model based on the concept of purchasing power parity. The monthly data from January 2010 to March 2019 are used for estimation. Two estimation techniques are employed in the study. The first one is Auto Regressive Distributed Lag Model (ARDL) approach with Bound test statistic. The other is ANN model which is non-linear model. From ARDL results, we can conclude that an increase relative output lead to baht appreciation of 0.692%. According to the relative important variables from ANN results, it shows that relative output is the most influential variables for Thai baht exchange rate. The result from ANN is consistent with the ARDL model. In addition, in terms of error estimation comparison between non-linear model as ANN and simple linear model, the result shows that ANN perform better than the linear model.