Font Size: a A A

Research On The Synthesis Method Of Literature Correlation Score

Posted on:2021-01-10Degree:MasterType:Thesis
Country:ChinaCandidate:M YanFull Text:PDF
GTID:2428330602472793Subject:Information Science
Abstract/Summary:PDF Full Text Request
Firstly,the literature degree of relationship is calculated by the common correlation coefficient method and the principal component method in this paper.When using correlation coefficient to calculate correlation score,the influencing factors of correlation score are analyzed,including correlation variable,weight function,parameters of the weight function,and the number of literature pairs.The effect of these factors on the relative distance and proportional relationship of variable weight,and the influence of these factors on the relative distance and proportional relationship of correlation score are analyzed.Secondly,the correlation score calculated by the correlation coefficient method and the principal component method is combined.In addition to the common principal component method,the similarity degree is also used to combine the two correlation scores,so that the similarity degree of the combined variables and the two correlation scores is the most consistent.When using similarity degree to combine the correlation scores,the combined variable is a linear combination of the two correlation scores,so to find the combined variable is to find the coefficient of two correlation scores in the linear combination,and the equation of similarity degree method is nonlinear.In addition to the equation,the condition is that the similarity degree is greater than zero,and the module of the vector corresponding to the merged variable is 1.Because of the non-linearity of the equation and some conditions,the paper uses the solver module of Excel to find the coefficients of two correlation scores in the linear combination,and substitutes the coefficients into the combined variables,which can calculate the combined variables of each pair of papers.
Keywords/Search Tags:degree of relationship, weight function, degree of similarity, solver module
PDF Full Text Request
Related items