摘要
先对论文水平以及评委这两个因素进行双因素方差分析,结果显示出了论文水平具有显著性差异,而评委的水平无显著性差异.进一步通过引入相关度的概念对残缺的评分矩阵进行补缺,建立了以相容性为基准的最优决策向量的数学模型,给出了该模型求解的迭代算法,对论文名次以及评委的水平进行了评判与逆判.通过计算机模拟,验证明了该算法的收敛性以及模型所建立的方法的优良性.
In this paper, firstly, bifactor variance analysis is made on two elements: the level of theses and judges. The result shows that the level of theses has a conspicuous difference, but the level of judges has not. Further, a conception of correlation degree is used to fill a vacancy to the fragmentary matrix of marking. According to that principlewe have presented a mathematical model of the best decision vectors based on compatibility and its iteration algorithm. Then judge and the reverse are both made on the rank of position of theses and the level of judges. By simulated test of computer, convergency of this algorithm and superiority of the methods made through the model have been verified.
出处
《湘潭大学自然科学学报》
CAS
CSCD
北大核心
2005年第2期39-43,共5页
Natural Science Journal of Xiangtan University
关键词
双因素方差分析
相关度
相容性
残缺数据
评判与逆判
bifactor variance analysis
correlation degree
compatibility
fragmentary data
judge and the reverse
