Using Multilayer TMDOM Model to Organize and Retrieve Textual Information

Infoglut that comes with the development of Internet-connected computers results in difficulties of knowledge management (KM). So making an efficient organization and retrieval of overwhelming amount of web resources ad hoc electronic document resources becomes more and more crucial. The Topic Map (TM), as an effective technology that is good at modeling and representing knowledge resources, is adopted in our study aiming to organize and retrieve textual information efficiently. In this paper, a multilayer Topic-Map-based model is proposed to organize the underlying concepts induced from domain document contents as the hierarchical structure by which the domain textual resources can then be linked to the related concepts in TM. Compared with the traditional statistics-based model, the searching process using TM-based model can be bounded in the narrow topic space, which leads to a high-quality retrieval results in terms of precision and recall. The experiment in the given domain is designed and implemented to evaluate the performance of the proposed model. The experimental results show that the efficient knowledge navigation can be made through building various associations between topics and textual resources.