Forecasting the wind speed process using higher order statistics and fuzzy systems

Abstract

This investigation has two main objectives. The first one is to propose a statistical method, based on the fourth order cumulants, to identify single-input single-output ‘SISO’, finite impulse response ‘FIR’ system using non gaussian input, zero mean and independent identically distributed signals. The second objective is to search, as an application, a model for forecasting the wind speed time series and to compare the obtained results with those obtained using the Takagi- Sugeno ‘TS’ fuzzy techniques. The prediction results obtained by the proposed method show that the sequences of generated values have the same statistical characteristics as those really observed and better than those obtained using ‘TS’ fuzzy systems. Additionally, the model developed on the basis of the statistical method fits well wind speed time series and can be used for forecasting purpose with an accuracy of 94 % and above.