Abstract: The paper analyses the relations between entities and activities based on enterprise subject databases, and proposes the affinitive degree and the polymeric degree for entities based on active-entity relations to create the polymeric degree matrix. According to the sizes of entity polymeric degrees, the method for dividing subject databases is presented to obtain the core entity pair, which is used to polymerize the same entity between the other second entity pair and the core one. The entity pair is calculated with weighted polymeric degrees to divide into some subject databases based on the core entity pairs. The example compares the calculation results of the affinitive degree matrix and the polymeric degree matrix, which shows that it is distinctly reflect the affinity and the trend among entities by creating subject databases based on the polymeric degree matrix.

Key words: subject database, entity, polymeric degree

摘要： 基于企业主题数据库分析了实体与活动的关系，提出了基于活动-实体联系的实体亲合度和聚合度，建立了聚合度矩阵。给出了主题数据库的划分方法，按照实体聚合度大小产生核心实体对，并以此核心来聚合其他与核心实体对有相同实体的次要实体对，进行加权聚合度计算，划分成多个以核心实体对为基础的主题数据库。实例比较了亲合度矩阵和聚合度矩阵产生的结果，表明通过聚合度矩阵建立主题数据库，能清楚地反映实体之间的相关性和趋向性。

关键词: 主题数据库, 实体, 聚合度

CLC Number: 

• TP311.131