Forex market forecasting using machine learning: Systematic Literature Review and meta-analysis

Abstract Background When you make a forex transaction, sell one currency and buy another. If the increases against sell, profit, do this through broker as retail trader on internet using platform known meta trader. Only 2% of traders can successfully predict movement in market, making it most challenging tasks. Machine learning its derivatives or hybrid models are becoming increasingly popular market forecasting, which is rapidly developing field. Objective While research community has looked into methodologies used by researchers to forecast there still need look how machine artificial intelligence approaches have been whether any areas that be improved allow for better predictions. Our objective give an overview their application FX market. Method This study provides Systematic Literature Review (SLR) algorithms forecasting. looks at publications were published between 2010 2021. A total 60 papers taken consideration. We them from two angles: I design evaluation techniques, (ii) meta-analysis performance utilizing metrics thus far. Results The results analysis suggest commonly utilized assessment MAE, RMSE, MAPE, MSE, with EURUSD being traded pair planet. LSTM Artificial Neural Network prediction. findings also point many unresolved concerns difficulties scientific should address future. Conclusion Based our findings, we believe area prediction room development. Researchers interested creating more advanced strategies might use open raised work input.