Experimental study of semantic contents mining on intra-university Enterprise Contents Management System for knowledge sharing

Abstract

We developed an Enterprise Contents Management System for an academic domain. The main feature of this system is its function for focusing searches in Web documents, utilizing human names and locations appearing in the documents as the search context. To realize this function, we adopted a standard text-mining algorithm for extracting proper nouns. We conducted an experimental study of this system against the existing digital contents of our university, and succeeded in efficiently obtaining suitable contents along the given contexts, which were obtained through previous searches. This experiment also suggested that our approach solves the general problem of finding an appropriate set of key words in a Web search. By performing this experiment, we confirmed that context mining is one of the most important technologies to be further developed in our effort to promote knowledge circulation through digital contents. © Springer-Verlag Berlin Heidelberg 2006.