基于密度聚类的学位论文质量评价方法研究

Quality evaluation on dissertation is an important means to measure the academic level of dissertations. The grouping evaluation method based on the set index weight is highly subjective and the evaluation scale is inconsistent, which leads to the difference in the evaluation results of the dissertations with similar quality. Based on the evaluation standards of some universities and the Ministry of Education, this paper constructs a set of evaluation index system of dissertation quality. Considering evaluator familiarity and scoring reliability, a scoring statistical method based on minimum relative difference is designed. In order to improve the stability and accuracy of the density clustering methods, an improved density clustering method based on two-stage search is proposed, and the method is used to evaluate the quality grade of dissertation and analyze the correlation of indexes. The case study shows that the clustering analysis of the overall quality of dissertation is helpful to find and eliminate the problems in the evaluation process, and to construct a reasonable evaluation index system. The proposed method also provides a new and effective way for dimension reduction of multi-dimensional data with category characteristics.