ABSTRACT
In the current chapter advanced short-term forecasting and trading is attempted by deploying artificial neural networks optimized by an adaptive genetic algorithm. The proposed evolutionary algorithm is an adaptive genetic algorithm which is used to optimize the structure and the parameters of two different types of neural networks: multilayer perceptron (MLPs) and wavelet neural networks (WNNs). Wavelet neural networks have been introduced as an alternative to MLPs to overcome their constraints, presenting more compact architecture and higher learning speed. The motivation of this chapter is to uncover novel hybrid methodologies which are able to acquire profitable short-term forecasting and trading strategies. The proposed models were applied to the task of forecasting and trading with the DJIE and the FTSE100 indexes to exclude liquidity problems from our study. The proposed hybrid techniques (aGA-MLP and aGA-WNN) were compared to some traditional techniques, either statistical, such as an autoregressive moving average model (ARMA), or technical, such as a moving average convergence/divergence model (MACD). The trading performance of all models is investigated in a forecast and trading simulation on our time series over the period 1997–2012. As it turns out, the aGA-WNN hybrid methodology does remarkably well and outperforms all other models in simple and more complex and elaborate trading simulation exercises.
