Correlation Analysis And Modeling Of Power Grid Planning Indicators Based On Copula Function

It’s very common that any of them in the index system of multiple indicators of change will cause the other indicators change phenomenon,and what all we could approach are covered by it is domains and fields.How to analyze and measure the correlation degree among the indicator variables is the premise of judging mutil-evaluation index system.Ignoring the correlation between indicators will affect the system’s reliability and economy,cause engineering calcul ation error,which affect the accuracy of decision-making.Power grid planning as area is an important part of national economy and social development,and there is the linear correlation index as well as nonlinear correlation index,Analysis simply on the basis of the linear correlation can not complete accurately express their correlation.In this paper,the correlation analysis and modeling method of power grid planning index based on Copula function model are proposed and the correlation measure is intr oduced.First,design the marginal distribution based on kernel density estimation algorithm,in view of the graphics method,existed mathod copulas model function have subjective and quantitative standard defect.So put forward the copulas connect function model of attribute recognition theory based on entropy weight is the preferred method;Secondly,the algorithm is designed and implemented to reduce the computational complexity.Finally,the corresponding program is written by Matlab,and the above algorit hms and models are studied and analyzed.Experiment example analysis shows that the method of paper selected copulas connect kernel function better characterize the related structures,accurately depict the end characteristics among all indicators,avoid t he traditional correlation analysis focus on the related degree of faults;Compared with the graph method,the algorithm of step estimation in the design and implementation of the thesis not only provides the quantitative evaluation method,but also the res ult is more objective and comprehensive,and has certain advantages.Compared with the traditional Pearson correlation coefficient method,this paper puts forward that the correlation analysis method based on copulas connect function,can more comprehensiv e related degree between the measure,contains more abundant relevant information.It provides strong support for the modeling and processing of power grid planning in the future,and has good practical significance.This paper conducts an in-depth study on the determination of edge distribution,selection of function model and parameter estimation in the construction of the Copula function model.The main work is as follows: 1.This paper designs the edge distribution estimation algorithm based on kernel density estimation,and implements the algorithm.With no need for the index distribution of prior knowledge,this algorithm only starts from the index data samples,after the analysis of the distribution features of the index data samples,kernel density estimation accuracy b y comparing with empirical distribution.2.An optimization method of the Copula function model based on the entropy weight attribute recognition theory is proposed,and an example analysis is carried out.The difference of Kendall rank correlation coefficient,the difference of Spearman rank correlation coefficient,Euclidean distance and maximum distance were used as evaluation indexes,and entropy weight attribute recognition theory was applied to grade five commonly used Copula models.This method shows some advantages as it performs more specific,more comprehensive,and more objective than the graph observation method.3.A two-stage step-by-step estimation algorithm is designed.The correlation parameters of five types of Copula func tions and Kendall rank correlation coefficient and Spearman rank correlation coefficient were calculated when this estimation algorithm is applied in the analysis of numerical examples.Based on these parameters,the density function graph and distribution function diagram of the Copula function model are drawn,which provides the conditions for the selection and evaluation of the model.