Sparse matrix computational techniques in concept decomposition matrix approximation

Abstract

Recently the concept decomposition based on document clustering strategies has drawn researchers' attention. These decompositions are obtained by taking the least-squares approximation onto the linear subspace spanned by all the concept vectors. In this chapter, a new class of numerical matrix computation methods has been developed in computing the approximate decomposition matrix in concept decomposition technique. These methods utilize the knowledge of matrix sparsity pattern techniques in preconditioning field. An important advantage of these approaches is that they are computationally more efficient, fast in computing the ranking vector and require much less memory than the least-squares based approach while maintaining retrieval accuracy. © Springer Science+Business Media B.V. 2009.