Comparative Analysis of Machine Learning Algorithms in Pair Currency Price Prediction in Forex Market, particularly in EURUSD and XAUUSD(GOLD)

Abstract

This thesis evaluates functions of machine learning algorithms to predict the price of EURUSD and XAUUSD, to help traders and financial institutions to make suitable decisions in using these algorithms. By comparing different algorithms like LSTM, ARIMA, SVM, and CNN, the main questions about accuracy of predictions and effects of some parameters are answered. Then basic concepts of time series analysis and financial data analysis are explained to show some information to understand price movements in markets. Evaluation metrics such as Mean Squared Error and Mean Absolute Error are used to evaluate model’s function. The steps of data preparation, model building, and evaluation are mentioned, then practical steps of algorithms are provided. By doing tests with EURUSD and AUUSD history data, the performance of models is evaluated. SVM was the most effective model in predicting both EURUSD and XAUUSD, after that LSTM, CNN, and ARIMA are on next ranks. In conclusion, this thesis provides an interesting look at the use of machine learning in the field of currency pair price prediction with the aim of helping to improve the skills of traders.