Abstract: This paper elaborates how to make use of hierarchical clustering to automatic divide fields combining the characteristic of literatures，mainly introduces the text cluster process and the key of techniques，in discipline domain realizes academic level appraisal in the author bunch，applies the h index to appraising author’s academic level，recommends the experts in specific field.It combines hierarchical clustering in the partition of literatures with the Web community together expanding the Web community’s concept.It applies the modifier h index to appraising author’s academic level fairly.

Key words: text mining, hierarchical cluster, community discovery, modifier h index

摘要： 结合文献资料特征利用层次聚类算法，实现了自动推荐学术领域专家系统。主要介绍了文本聚类的过程及关键技术;完成对学科领域内作者簇内学术水平的评价，将h指数应用到评价作者的学术水平之中，推荐出该领域中的专家群。将文献资源的层次聚类划分与专家社区发现结合在一起，采用修正h指数公正地评价了作者的学术水平。

关键词: 文本挖掘, 层次聚类, 社区发现, 修正h指数