Foreign Currency Exchange Rate Prediction Using Bidirectional Long Short Term Memory

Abstract

This chapter aims to predict the foreign currency exchange rate over twenty-two different currencies based on the US dollar. This chapter proposes three machine learning algorithms, such as ridge regression, lasso regression, decision tree, and a deep learning algorithm named Bi-directional Long Short-Term Memory (Bi-LSTM) to predict the foreign currency exchange rate. Technical analysis of foreign currency exchange is also discussed in this chapter. The authors use mean absolute error (MAE), mean square error (MSE), root mean squared error (RMSE), and mean absolute percentage error (MAPE) to measure the performance of the algorithms. Empirical findings indicate that overall performance of all the algorithms is satisfactory, but Bi-LSTM performs better than others. This study is beneficial for the stakeholders in setting a range of strategies for the foreign exchange market.