Skip to Main Content
In the application of neural network model for short term load prediction, main problems are over many training samples, long training time and low convergence speed. For representative training samples, an ant colony clustering model based on Elman neural network was proposed in this paper. First, historical load data were pre-processed by using ant colony clustering method. The clustered data were chosen as training samples for the network. The objects are to make the input samples representative, decrease training time, increase convergence speed and improve prediction accuracy. Based on daily load data of one electric power plant, this model can obtained more accurate prediction results.