基于语义标注的文本聚类算法研究

摘要

How to improve the automatic creation of metadata quality has become more and more widely recognized problems. The rapid growth of data resource has increased the importance of automatic metadata creation. At present, there are many automatic creation of metadata document research. They use the same principle, but the details are slightly different. Based on the collaborative annotation, a complementary approach is proposd, based on collaborative annotation information to reflect the opinions about the content in a document from different readers' view. And then a method of automatically constructing metadata based on Self-organizing map algorithm is addressed. The method can improve the quality of automatically-generated metadata.%为了改变元数据创建还处于手动、半手动的现状,要提高Web页元数据的精确度.首先基于语义标注理念,以群体共享的形式对Web页语义信息进行标注.生成过程中利用自动聚类算法,侧重于分析标注者不同特点对提高元数据质量所起的作用,提出了一种自动生成元数据方案.最后,通过实验证明,元数据结果会因标注者特点的不同而异,并且当标注者影响超过了临界值会提高元数据的质量.