基于等价性检验和特征聚类的《红楼梦》作者分析

An Analysis of the Authors of A Dream of Red Mansions Based on Equivalence Checking and Feature Clustering

摘要: 引入等价性检验模型，选取“红”、“玉”二字统计频数，计算检验统计量U与p值。根据U检验值与概率对照表，初步得出前80章与后40章存在差异，并非一人所著。同时选取K均值聚类与凝聚聚类，根据词频聚类出多种情况。结果表明，《红楼梦》全书使用词频均存在着差异，其作者不止一人。

Abstract: The equivalence checking model is introduced to calculate the test statistics U and p values by se-lecting the statistical frequency of “red” and “jade”. According to U-tests and probability compara-tive table, differences between the first 80 chapters and the last 40 chapters are preliminarily concluded. Many cases are clustered by word frequency with K-means clustering and agglomerative clustering. The results show that there are differences in word frequency used in A Dream of Red Mansions, and there is more than one author.