A method of Chinese document expression for extracting formal context

As a kind of data model, a formal context must be extracted from some actual data sources such as documents. For case of unstructured Chinese document, it is the first question to decide how to express the document. Vector space model (VSM) which is the dominant model of document expression now takes a single word as a feature item, so that neglects the lexical semantic relationship between words in a natural language, thereby it cannot show the semantic information implied in the documents. This paper discusses an improved method which is to take Hownet as knowledge base, to establish the concept vector space of Chinese document by using the set of similar word set to replace the single feature word in VSM. This method is convenient to extract a formal context according to the threshold made by users, and how to apply it is illustrated by an example.