Selection of orthogonal investment portfolio using evolino RNN trading model

Abstract

Investing in financial market require the reliable predicting of expecting returns, assessment of risk and reliability. Principle of portfolio orthogonality was using to reduce the risk of the investment. An artificial intelligence system may reveal new opportunities for using this principle. Prediction of recurrent neural networks ensemble is stochastically informative distribution, which is helpful for portfolio selection. Shape and parameters of distribution influence decision making in currency market. Assessment of portfolio riskiness, finding most orthogonal elements of portfolio, influence better results for trading in real market.