An Application of Meta Search Agent System Based on Semantized Tags for Enhanced Web Searching

Web searching techniques have been investigated and implemented in many aspects. Particularly, in case of personalization, more important issue is how to manipulate the results retrieved from search engines for better user understandability and satisfaction. Such manipulation processes are i) ranking the results in accordance with user relevance, and ii) exchanging the results between users who have similar tastes. Thus, our work has been mainly focusing on relevance-based ranking mechanism as well as sharing schemes for the results retrieved from heterogeneous web information sources. In this paper, we propose a hybrid model for meta search agent systems with three main functionalities, i.e., i) URL filtering method for preprocessing, ii) tag-based information conceptualization scheme for ranking, and iii) ontology-based standardization scheme for sharing. It means that the proposed meta search agent model exploits semantized tags to formalize and share heterogeneous information obtained from multiple search engines and to finally maintain the shared information. Within the tag-based information space, a conceptual distance between retrieval interest and search results can be efficiently computed. By conducting some experimentations, we have shown the semantized tag model can conceptualize the retrieved results, and make them sharable. We also compare performance of the proposed system with hyperlink-based methodologies.