Knowledge Management System Based on Latent Semantic Analysis

The base of described knowledge management system is a semantic space that is an outcome of latent semantic analysis performed on certain set of documents. In such a space, defined with semantic coordinates, every document is projected as a vector. Ontology is used to describe hierarchy and relations among categories. Additionally, semantic vector is assigned to every category. Assignment of some document to a certain category is defined with proximity of their semantic vectors. Learning and validity of such system is guaranteed with loopback inputs provided by category moderator.