Semantic text classification with tensor space model-based naïve Bayes

This paper presents a semantic naïve Bayes classification technique that is based upon our tensor space model for text representation. In our work, each of Wikipedia articles is defined as a single concept, and a document is represented as a 2
^nd
-order tensor. Our method expands the conventional naïve Bayes by incorporating the semantic concept features into term feature statistics under the tensor-space model. Through extensive experiments using three popular document collections, we prove that the proposed method significantly outperforms the conventional naïve Bayes. Surprisingly, the classification performance amounts to almost 100% in terms of F1-measures when using Reuters-21578 and 20Newsgroups document collections.