Lexical Semantic SLVM for Semi-structured Document Classification

Structured Link Vector Model (SLVM) and its improved model depend on statistical term measures to implement semi-structured document representation. As a result, they ignore the lexical semantic contents of terms and the distilled mutual information, leading to text classification errors. This work proposed a document representation method, WordNet-based lexical semantic SLVM, to solve the problem. Using WordNet, this method constructed a data structure to characterize lexical semantic contents, and adjusted EM modeling to disambiguate word stems. Then, feature matrix for document representation was built in the lexical-semantic feature space of semi-structured document, and applied to NWKNN classification algorithm. On categorized dataset of Wikipedia XML, the experimental results show that the feature matrix of our method performs F1 measure better than original SLVM and frequent sub-tree SLVM based on TF-IDF.