Research of short-term load forecasting based on combined grey neural network and phase space reconstruction

According to the characteristics of grey theory, G-P algorithm of phase space reconstruction and artificial neural network (ANN), a combined algorithm (G-G-NN) is proposed. The original time series is transformed by accumulated generating of grey prediction and G-P algorithm of phase space reconstruction. When a regular time series phase space is generated, neural network is adopted to forecast. The practical example indicated that the algorithm is verified.