Stochastic and genetic neural network combinations in trading and hybrid time-varying leverage effects

摘要: Abstract The motivation of this paper is 3-fold. Firstly, we apply a Multi-Layer Perceptron (MLP), Recurrent Neural Network (RNN) and Psi-Sigma (PSN) architecture in forecasting trading exercise on the EUR/USD, EUR/GBP EUR/CHF exchange rates explore utility Kalman Filter, Genetic Programming (GP) Support Vector Regression (SVR) algorithms as combination techniques. Secondly, introduce hybrid leverage factor based volatility forecasts market shocks study if its application improves performance our models. Thirdly, specialized loss function for Networks (NNs) financial applications. In terms results, PSN from individual SVR forecast techniques outperform their benchmarks statistical accuracy efficiency. We also note that strategy successful, it increased most models, while NNs seems promising.