A Long-Term Forecast Method for the Investment Demand of Power Grid Based on Linear Regression and Error Correction of Variable

Abstract

In the long-term forecast for the investment demand of power grid, the error of independent variable will result in a large error between forecast result and actual result. In this paper, the correction parameters of factors which have an influence on investment demand of power grid is introduced and the correction model of long-term forecast is built. Firstly, through the correlation analysis and the principle comment analysis (PCA), the variables which the model needs are selected. Secondly, the long-term forecast model based on linear regression is built and the variable coefficient is calculated through least square method. Thirdly, after analyzing historical data of forecast value and actual value, correction model of long-term forecast is built through introducing correction parameters of each variables. Finally, compared with the calculation results which get from correction model and the long-term forecast model, it is proved that the error of correction model is smaller.