Text categorization using topic model and ontology networks

Text categorization based on pre-defined document categories is one of the most crucial tasks in text mining applications in recent decades. Successful text categorization highly relies on the text representations generated from documents. In this paper, an innovative text categorization model, VSM_WN_TM, is presented. VSM_WN_TM is a special Vector Space Model (VSM) that incorporates word frequencies, ontology networks and latent semantic information. Unlike the traditional text representation using only Bag-of-words (BOW) features, it incorporates semantic and syntactic relationship among words such as synonymy, co-occurrence and context, with the purpose of providing more inclusive and accurate text representation. Support Vector Machine is used as document classifier, and the proposed system is evaluated on three publicly available datasets and one domain-specific dataset. Experiment result shows that our approach significantly improves text classification by outperforming approaches such as using only latent features and traditional VSM approaches. Keywords— Text Categorization, vector space model, ontology network, topic model, support vector machine.