000012 Natarajan A M;Premalatha K (NO, Kongu Engineering College, Perundurai-638 052, Email: email@example.com) : Three dimensional concept-based indexing for XML documents. Int J Syst Cybernetics Inf 2007, (Jul), 40-5.
In concept based indexing the documents are viewed as collection of concepts and indexed according to the meaning rather than terms. This will require construction of concept hierarchy which consists of collection of concepts and their interrelationships. The proposed system uses concept hierarchy to extract concepts from the XML document. The semantic importance of each term present in the concept is calculated and it is represented in 3D Lattice Cube. But this one is sparse and not efficient in terms of space and access time. To improve the performance of the information retrieval, K-Means clustering is employed to generate several smaller lattice cubes that are not sparse. Now the queries can be directed to appropriate cluster and also querying is faster in a smaller lattice cubes.
11 illus, 1 table, 15 ref