Constructing Maintainable Semantic Relation Network from Ambiguous Concepts in Web Content

Abstract

The semantic network is a form of knowledge that represents various relationships between concepts with ambiguity. The knowledge can be employed to identify semantically related objects. It helps, for example, a recommender system to generate effective recommendations to the users. We propose to study a new semantic network, namely, the Concept Relation Network (CRN) , which is efficiently constructed and maintained using existing web search engines. CRN tackles the uncertainty and dynamics of web content, and thus is optimized for many important web applications, such as social networks and search engines. It is a large semantic network for the collection, analysis, and interpretation of web content, and serves as a cornerstone for applications such as web search engines, recommendation systems, and social networks that can benefit from a large-scale knowledge base. In this article, we present two applications for CRN: (1) search engine and web analytic and (2) semantic information retrieval. Experimental results show that CRN effectively enhances these applications by considering the heterogenous and polysemous nature of web content.