Abstract
Abstract Recently, due to the large amount of electronic information over the Internet, to efficiently retrieve and filter information that meets the user requirements has become an essential issue for enterprise document/knowledge management. In addition, in order to assist users quickly acquire the critical information in the documents, keyword extraction of each document is necessary. Traditionally, document keywords are determined manually and much expertise and time are required. In this research, algorithms for automatic keyword correlation analysis and keyword extraction are proposed. The keyword correlation is analyzed based on the word frequencies and locations of documents in the document repository. Based on the keyword correlation, a recursive approach for document keyword extraction is developed to effectively determine the critical and representative concepts of the target documents. A web-based system for document content management is established and a case study is provided to valuate the proposed model. The proposed approach is not domain-oriented and thus can be applied in different applications. This research is to provide a feasible solution for enterprise knowledge accumulation and reuse in the collaboration networks.