Research On Forecasting Model Of Power Grid Investment Demand Based On Data Mining

As an important implementation subject of the national energy strategy,the power grid undertakes the important responsibility of optimizing the allocation of energy resources and serving the economic and social development.With the rapid development of China,s national economy and the increasing demand for electricity in the whole society,the investment demand of China's power grid is increasing with the vigorous promotion of the "13th Five-Year plan".Accurate and effective forecasting for investment demand of power grid can not only help make overall planning to rationally arrange investment in power grid construction and reduce capital costs and economic risks,but also effectively improve the operation of power grid enterprises,which plays a vital role in ensuring the steady development of power grid enterprises and promoting the power grid investment planning and construction process.Firstly,this paper combs and summarizes the research trends of domestic and foreign scholars on the grid investment demand forecasting,data mining technology and forecasting technology,and the current situation of China's power grid investment demand and the commonly used forecasting methods for power grid investment demand are described iin detail.Then,on the basis of reading a large number of literatures and consulting experts,according to the selection principle of influencing factors,the influencing factors of power grid investment demand are analyzed and selected preliminarily from four aspects:macro-economy,power demand,power grid scale and power grid benefit.At the same time,the combination algorithm of grey relational analysis and kernel principal component analysis(GRA-KPCA)is used for identifying the key influencing factors of grid investment demand,and the effective input vectors of the forecasting model are obtained.Next,this paper proposes a model based on intelligent mining algorithm of support vector machine optimized by differential evolution algorithm and modified grey wolf optimization algorithm(DE-MGWO-SVM)for power grid investment demand forecasting,which combines the three algorithms of differential evolution,grey wolf optimization and support vector machine organically for secondary optimization of forecasting algorithms.The improved combination forecasting model can realize the complementary advantages between different algorithms,and can greatly improve the prediction accuracy.In addition,the applicability of the model is analysed in this paper,and compared with the commonly used methods of power grid investment demand forecasting,such as regression analysis forecasting method,time series forecasting method,grey forecasting method and artificial neural network forecasting method.Finally,this paper takes the power grid of a certain area in East China as an example to make an empirical study of the proposed model.At the same time,the models without factor identifying and screening,the models without multiple parameters optimization and the traditional forecasting models are used to forecast the same sample,and the prediction results of different models are compared and analyzed,which proves that the power grid investment demand forecasting model based on DE-MGWO-SVM has high forecasting accuracy.In this paper,a novel power grid investment demand forecasting model based on data mining is proposed taking data mining as the means and taking the power grid investment demand forecasting model as the research object.Through empirical research,the effectiveness,applicability and superiority of the proposed model for power grid investment demand forecasting are proved,which provides a new method for power grid investment demand forecasting.