Chaotic Load Series Forecasting Based on MPMR

In this paper, minimax probability machine regression (MPMR) is proposed for chaotic load time series global prediction. In MPMR, regression function maximizes the minimum probability that future predication will be within an epsiv to the true regression function. After the theory of MPMR explored, and the chaotic property of the load series from a certain power system verified, one-day ahead predictions for 24 hours points next day were done with MPMR. The results demonstrated that MPMP had satisfactory prediction efficiency. Kernel function shape parameter and regression tube value influences the MPMR-based system performance. In experiments, cross validation was used to select the two parameters