残缺数据的论文名次及评委水平的评判与逆判

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.

Judge and the Reverse for the Rank of Position of Theses and the Level of Judges of Fragmentary Data

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.