Research On A Hybrid Wavelet Transform And Adaptive Files Optimization Algorithm Of Svm In Short Term Load Forecasting

Contemporary society can not be separated from the power,because the power industry is not only leading the economic development,but also to promote national economic growth,improve the quality of life of the core pillars.The main measures to protect the safe and reliable operation of power system are short term power load forecasting,which is related to the quality of life of all walks of life and residents.Only the high accuracy prediction can be used as the key index of the security dispatch,the economic operation and the reasonable planning to improve the reliability of power supply.It is necessary to deal with the huge data.This paper reviews the related methods of load forecasting,data mining,and according to the load characteristics of data,comparison and selection of correlation method,improved method and lay a solid foundation for the research of short-term load forecasting is more profound.Because of the advantage of the support vector machine,it is chosen as the main method of prediction,and the superiority of the method is proved by experiments.In order to optimize the accuracy of the model,use a variety of methods and improve optimization algorithm: firstly,The wavelet analysis is used to deal with the electric load data,which makes the randomness and regularity of the data more obvious;second,the support vector machine parameters are directly related to the accuracy,but the means of selection is scarce,in order to solve this problem,the mutation algorithm is introduced to select the appropriate parameters.Finally,using the historical load data and meteorological data in a region of Henan Province to analyze the prediction model.The results show that the proposed method not only has good convergence and fast training speed,but also has high practical significance.