Short-term power load forecasting based on combined kernel Gaussian process hybrid model

As one of the countries with most energy consumption in world, electricity accounts for a large proportion supply our country. According to national basic policy conservation and emission reduction, it is urgent realize intelligent distribution management by prediction. Due complex nature load sequences, traditional model predicts poor results. kernel-based machine learning model, Gaussian Process Mixing (GPM) has high predictive accuracy, can multi-modal prediction output confidence intervals. However, GPM often uses single kernel function, effect not optimal. Therefore, this paper will combine variety existing build new kernel, use sequence In experiments, characteristics sequences are first analyzed, then made based on optimal hybrid function constructed compared model. The results show that only superior but also some models such as ridge regression, regression GP.