Evaluation of the portfolio performance indicators, using Evolino RNN trading model

Abstract

The conditions for investment depend on correct forecasts of financial markets parameters received by artificial intelligence based on currency trading models, where AI was used as probabilistic forecasting tool. Improvements of predictions probability directly influence the investment portfolio performance indicators and their inter-relationship. Newly developed Evolino Recurrent Neural Network (RNN) based on forecasting model was used for study of influence of currency exchange rate forecasts on investment decisions related with reception and choice of investment strategy. Results of portfolio performance indicators and portfolio riskiness are presented.