Wind Speed Forecasting Using Empirical Mode Decomposition and Regularized ELANFIS

This paper proposes a novel hybrid method to predict the wind speed using empirical mode decomposition (EMD) and regularized extreme learning adaptive neuro-fuzzy inference system (RELANFIS). RELANFIS combines the learning capability of conventional adaptive neuro-fuzzy inference system (ANFIS) and faster computational speed of extreme learning machine (ELM) algorithm. EMD decomposes original wind speed data into finite IMFs and one residue. Then, each decomposed data series is predicted using RELANFIS model. Final prediction is obtained by the summation outputs of all RELANFIS sub models. Performance comparison with several popular EMD wind speed prediction methods shows that hybridization of EMD with RELANFIS gives the best prediction results.