A novel algebra to articulate feature in text dimension reduction

Abstract

Challenges in text mining arise from multi-corpus and high dimensionality involving natural language. Features from datasets needs to be composed or articulated. This paper aims to develop a new approach to align text feature using ontology, which can form the base of text dimension reduction. Firstly, a novel text feature graph is defined, based on which we can do articulation. Secondly, an algebra system is proposed for text feature graph computing. Finally, an instance is demonstrated to show the efficiency and accuracy of the proposed approach.