Ranking Evaluation Method And Its Application

With the rapid development of information technology,people are faced with more and more decision-making problems in the era of big data.How to select the best individual in a group of groups is the central issue in ranking research.Benefit from the promotion of related applications,such as political elections,world university rankings,talent selection and sports events rankings,ranked issues in the past few decades has been extensively studied,people are committed to ranking and use of rankings,but neglected the evaluation of rankings.In practice,in order to improve the objectivity and fairness of rankings,the same group of objects often have multiple rankings,for example,university rankings published by different institutions and rankings given by different judges in competitions.How to evaluate the rankings according to the information given by multiple rankings is called ranking evaluation.Rank evaluation can measure the objectivity and fairness of ranking,identify the "bad guys" in multiple rankings,guide rankings improvement,and obtain more rankings resultsBased on the ranking theory,this article comprehensively uses multidisciplinary fields such as graph theory,probability theory,mathematical statistics and computer simulation to discuss how to evaluate the ranking's fairness Degree "and" how to evaluate the degree of stability of rankings ",and deeply studied the modeling,analysis and application of ranking evaluation methods.The main research work and innovation of this paper are as follows:(1)An iterative ranking method is proposedAiming at the shortcomings of the current ranking aggregation method,an iterative ranking aggregation method is proposed.By iteratively updating the weight of a single list,the contribution of the good ranking list to the aggregate ranking increases,while the contribution of the poor ranking list decreases,the final aggregate list closer to the true rankings,radically improved the traditional rank aggregation method.In order to evaluate the effectiveness of the ranking aggregation method,a computer simulation technique was used to establish a ranking process simulation model.It was found that the iterative ranking method can reduce the number of inconsistencies in the ranking and provide more accurate rankings.(2)Three ranking evaluation methods are proposedAiming at the shortcomings of the current rankings evaluation methods,this paper proposes rankings based on outliers,fairness-based and stability-based approaches from the perspective of rankings anomaly,fairness and stability.The index of outlier of single ranking object is given,and the outlier model of ranking list is proposed.The definition of group and group quality is given,and the model of ranking is put forward.The definition of stability is given.A model for ranking stability is presented.In order to show the result better,the above three evaluation methods are respectively visualized by hotspot,bubble chart and line chart.(3)Study of the world university list evaluationIn this study,we apply rankings evaluation methods to university rankings.According to the current list of world universities,we use the above methods to measure and evaluate the rankings of anomalies in five well-known rankings,and whether the rankings of universities in different regions are fair and objective.From the time dimension,the list is stable enough and found that the list of QS rankings and other charts vary widely,the location of the issuing list of institutions will have some local preferential treatment,but in the time dimension of the analysis,the world The university list does not provide a ranking that is stable enough to change.The evaluation of university rankings helps to improve the rankings of existing university rankings and is instructive to university rankings.