Research on Ontology-Based Text Representation of Vector Space Model

In traditional Vector Space Model (VSM) the TF*IDF method is widely used to adjust the weight of terms in text mining. However TF*EDF can not represent the semantic information of text by neglecting the semantic relevance between terms. In this paper, an improved ontology-based VSM is presented, in which the ontology-based term similarity is used to readjust the weight of semantically related terms. The experimental results show that the improved VSM can perform more accurately than the traditional VSM in the calculation of the term weights.