IV. CONCLUSIONS This paper presented the development of a novel fuzzy neural network model and validated its prediction on the short
term electric load forecasting of the Power System of the Greek Island of Crete. In the proposed scheme, a two-stages clustering algorithm to determine the rules, number of fuzzy sets, and initial values of the parameters (centers and widths) of the fuzzy membership functions has been employed. The forecasting error statistical results, corresponding to the minimum and maximum load time-series, indicate that the proposed load forecasting model provides more accurate forecasts, compared to conventional neural networks models such as MLP and RBF. In a future work, the present approach will be enhanced by using additional load and weather information such as illuminations level, temperature, humidity, wind direction and industrial load.