Abstract
Predicting the electricity market price has an important role in the future decision-makings in energy consumption. In this study, three intelligent combined methods: Elman neural network-Genetic Algorithm (El-GA), Elman neural network-Imperialist Competitive Algorithm (El-ICA) and Elman neural network-Particle Swarm Optimization (El-PSO) are proposed to predict the short-term price of electricity, and are evaluated by different error criteria. The data, which are used in this paper, are daily and related to a one-year period (2006), in the USA. The neural network input in this research is different ahead in time periods of electricity price data. Also, in order to have a better evaluation of the performance of the suggested combined methods, the variance analysis (ANOVA) method is employed; and then, based on the mean and distribution of MMAPE error criteria the appropriate method is selected. In the present study, El-PSO combined method has a better performance as compared to other methods such as El-ICA, El-GA, MLP and SVM.