Abstrait

Classification of Documents in E-Learning Using Multidimensional Latent Semantic Analysis

R.Archana, M.Ravichandran

In this paper we consider the problem of dimensionality reduction techniques. Two techniques such as Independent Component analysis (ICA) and multidimensional latent semantic analysis (MDLSA) are proposed. A new document analysis method named multidimensional latent semantic analysis (MDLSA) which resolves the problem of in-depth document analysis, mines local information from a document efficiently with respect to term associations and spatial distributions. The MDLSA first partitions each document into paragraphs and later builds a term ―affinity‖ graph. Each element of this graph represents the frequency of term co-occurrence in a paragraph. We then use Independent Component Analysis (ICA) which finds a linear representation of nongaussian data such that the components are statistically independent. Thus these two techniques are examined in retrieving and classifying the e-learning documents. It is also proven by experimental verifications that the proposed technique outperforms current algorithms with respect to accuracy and computational efficiency.