Output details
11 - Computer Science and Informatics
University of Westminster
A clustering-based fuzzy-wavelet neural network model for short-term load forecasting
<02>Originality: This paper presents a novel clustering-based hybrid intelligent system that utilise multidimensional Gaussian functions and a modified Expectation Maximization algorithm. Two schemes have been developed based on this concept, one fuzzy-wavelet neural network (FWNN) and one neuro-fuzzy network. A hybrid learning algorithm has been applied to both schemes.
Significance: The forecasting models validated their accuracy on the prediction on the load forecasting of the Power System of the island of Crete.
Rigour: FWNN was successfully compared against other models. Results published in a highly peer-reviewed journal, currently ranked in the 2nd position at the ISI Web of Knowledge (AI).