Forecast Of Electricity Price Based On Data Analysis And The Model Of Improved DE-SVM

The decision-making of power generation enterprises,power supply enterprises,and power consumers can be affected by forecasting the price of electricity.The data combined with the electricity price and the feature is huge.There are many irrelevant samples and features in the data,which often lead to low forecasting accuracy and high time-cost.Therefore,this paper conducts data analysis and the research of price forecasting and constructs a forecasting framework,which reduces the dimension of data from sample features and uses improved DE-SVM to forecast the electricity price.Finally,a small quantity of data is selected to achieve accurate forecasting while reducing the time-cost.The main conclusions of this paper are as follows:(1)The method and idea of the data analysis and forecasting model are analyzed.Given the problem of forecasting accuracy and time-cost,the improvement direction of data analysis and price forecasting is discussed.(2)For the model of the data analysis: Firstly,aiming at the problem of a set of sample sets being over-extracted data,considering the periodicity and time relevance of electricity price,a data processing method based on feature classification is proposed.Secondly,given the problems of the forecast accuracy and time-cost of the forecasting model of the electricity price caused by data redundancy,combined with gray correlation analysis(GCA)and principal component analysis,it is proposed to clean up the massive data from samples and features.Also,the GCA is improved,and a multi-objective reference sequence is established.Finally,an important sample set is obtained,while the single-point forecasting problem is transformed into a time-period forecasting problem.(3)For the forecasting model of the electricity price: The model of electricity price forecasting is composed of a differential evolution algorithm and a model of support vector machine(DE-SVM).Considering the influence of outliers on the DE-SVM forecasting model,this paper proposes a forecasting model of the electricity price based on improved DE-SVM in the error feedback unit,the root mean square error and the mean absolute error are introduced,and the exponential function is applied to combine these two indicators to further improve the forecasting accuracy.(4)A case study of the New England electricity market is presented.This framework is compared with other models.Through the case analysis of the model of the data analysis and the model of electricity price forecasting,the feasibility and effectiveness of the framework are verified.The framework can effectively improve the forecasting accuracy and reduce the time-cost.It provides a new idea for data analysis and electricity price forecast.