Abstract
ABSTRACT Due to popularity of the information technology, the electronic documents within enterprises and organizations accumulate quickly and drastically. To automatically classify documents is the typical issue for enterprise knowledge management and services. Currently, most of the automatic document classification methodologies utilize the document keywords for determination of document category. Furthermore, most previous research about keyword extraction focuses mainly on improvement of the extraction methodology and the characteristics of document keywords are rarely taken into consideration. Therefore, in this research, the key characteristics of document keywords (e.g., the frequency and location) are analyzed and the results can be applied as the criteria for automatic keyword extraction. On the other hand, in order to enhance the accuracy of automatic document classification, an automatic document classification approach is also proposed on the basis of the document structure and document provider. In addition to the keyword analysis and document classification models, a web-based prototype system is developed for automatic document classification. The effectiveness of the developed system is evaluated via a e-News management case. This research attempts to explore applicable keyword characteristics and document classification mechanism so that the goal of automatic enterprise knowledge management can be realized. Keywords: Document Classification, Keyword Extraction, Knowledge Management, Information Retrieval.