篇名 | Electricity Sales Prediction Model of Electricity Market of the Non-linear Regression Combination with the Optimal Weight |
---|---|
卷期 | 30:5 |
作者 | Zhao-Yang Qu 、 Shu-Ya Xie 、 Rong-qiang Feng 、 Nan Qu 、 Hong-bo Lv 、 Zuo-Wei Chi 、 Chong Qu |
頁次 | 075-087 |
關鍵字 | cuckoo algorithm 、 electricity power market 、 least squares support vector machine 、 RBF neural network 、 EI 、 MEDLINE 、 Scopus |
出刊日期 | 201910 |
DOI | 10.3966/199115992019103005006 |
The model to estimate electric sales in electric market of the optimal weighting nonlinear regression combination is brought forth to improve the accuracy of electricity sales forecast. Firstly, the least square method is used to estimate the nonlinear regression equation in order to obtain the result sets of the prediction sequence. Secondly, the design method of minimizing the calculation of fitting variance and weighting com-bination to calculate the residual sequence and weighting of the Radial Basis Function (RBF) neural network and the least squares vector machine, finally the result sets of residual sequence concluded. At the end, a nonlinear decreasing strategy of inertia weight is introduced on the basis of cuckoo algorithm to optimize the extended search space of model parameters and the model solution. The simulation results show that the model can accurately and comprehensively reflect the law of electric sales and improve the accuracy of sales prediction.