Modelling the Currency Exchange Rates Using Support Vector Regression

In this work, we modelled the exchange rates using Support Vector Regression (SVR). The currency selected in this work was Turkish Lira (TRY) and compared the rates between United States Dollars (USD), Euro (EUR), and British Pounds (GBP). The modelling was done for the period of January 2018 till October 2019 data, with the aim to investigate three different issues. The difference in using ν-regression and e–regression was investigated. The effect of using different kernel functions was examined. Also, the effect of using multiple different exchange rates together (multi-rate) compared with modeling only one rate (single-rate) was explored. The results showed very successful modelling of the currencies producing very accurate results using the ν-regression modelling with radial kernel and using multi-rate modelling.