摘要
电网规划决策中,忽略指标间的相关性将影响整个电网规划的可靠性和经济性,引起工程计算的误差,进而影响决策的准确性。电网规划指标间不仅存在线性相关还存在非线性相关,单纯以线性相关为基础的分析研究不能完整精确地表达其相关性。为此,提出基于Copula函数的电网规划指标的相关性分析及建模方法,并引入了相关性测度。首先采用核密度估计法确定变量的边缘分布及图形观测法选取Copula函数,并利用极大似然法进行参数估计。然后以经验Copula函数为基础计算欧氏距离值,评价二元正态Copula和二元t-Copula两模型的优劣。最后,以某省电网规划方案的数据为样本进行实例验证,结果表明所提方法构造的Copula函数模型的有效性,不仅能有效刻画指标间的相关结构和尾部相关性,而且能更好地反映指标间的秩相关性。
The correlation between indexes will affect the reliability and economy of the whole power grid planning in the decision of power grid planning,which leads to the error of the engineering calculation,and misleads the accuracy of decision.The linear correlation and nonlinear correlation exist in the power grid planning indexes.The analysis and research only based on a linear correlation can′t express the correlation between indexes completely and precisely.Therefore,the correlation analysis and modeling method of power grid planning indexes based on Copula function is proposed,and the correlation measure is introduced.The kernel density estimation method is used to determine the marginal distribution of the variables.The graph observation method is used to select the Copula function,which is estimated with the maximum likelihood method.The empirical Copula function is adopted to calculate the Euclidean distance,and evaluate the merits of bivariate normal Copula and bivariate t?Copula models.The data of a certain provincial power grid planning scheme is taken as an example for instance verification.The results show that the proposed method is effective to construct the Copula function model,can not only depict the correlation structure and tail correlation between indexes effectively,but also better reflect the rank correlation between indexes.
作者
丁家满
唐渐
王清心
杜奕
DING Jiaman;TANG Jian;WANG Qingxin;DU Yi(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)
出处
《现代电子技术》
北大核心
2018年第17期95-101,共7页
Modern Electronics Technique
基金
国家自然科学基金(51467007)
国家自然科学基金(51365020)~~
