Exchange-rate-predictions

Foreign exchange rate is the rate at which one currency is exchanged for the other. Usually all currencies are compared against USD as a standard. Those exchange rates vary every day based on multiple factors. They are modeled as financial time series problems and still pose a major challenge to many professionals who try to forecast it accurately. Their inherent nature of being noisy and volatile makes is difficult for us to fit it in a form. Multiple techniques were contrived and applied to make their study easier. This research evaluates multiple methods like SVM, ARIMA, ANN, and RNN to forecast those exchange rates and compares their performances against one another. We find that LSTM performs best among all the techniques in non linear time series forecasting.