电力负荷预测线性相关方程的最佳权重加权最小二乘法

The Weighted Least-squares Method with Optimal Weight for Linear Dependent Equation of Power Comsuption Forecast

针对电力系统负荷预测的一元线性回归模型,提出了基于加权最小二乘法建立相关方程的最佳权重法.与传统的最小二乘回归法相比较,该方法综合考虑x及y方向误差确定相关方程,所求相关方程不会因坐标系的选取而改变,结果较为客观,可作为电力负荷预测的一种新的补充方法.

For applying to the linear regression model on power comsuption forecast, the weighted leastsquares method with optimal weight is proposed.It is different from the traditional leastsquares method that the errors of x and y directions are comprehensively considered during determining the dependent equation in the new method,so that the dependent equation is unchangeable with the selected direction and the result is more objective. The new method may be considered as a supplementary way for forecasting the power consumption.