A Fuzzy-Rough Set Based Semantic Similarity Measure Between Cross-Lingual Documents

As cross-lingual information retrieval attracts increasing attention, tools that measure cross-lingual document similarity become desirable. Since the way that people convey thoughts at the abstract concept level makes little, if any, difference in the languages they use, it is possible to measure semantic similarity between different lingual documents based on the concepts conveyed by the documents. In this paper, a novel fuzzy rough set based method for measurement of semantic similarity between cross lingual (Chinese and English) documents is proposed. Aided by a bilingual dictionary and Wordnet, translation is processed like word sense disambiguation and all the distilled senses are used to construct a fuzzy approximation space using a fuzzy partition algorithm. In the fuzzy approximation space documents are approximated by their fuzzy upper and lower approximations and the similarity measure is defined accordingly. The upper and lower approximations correspond to the slack and tight extent of the concepts in their associated document. This method makes possible to distinguish among the documents whose original texts seem not similar but conveyed concepts are similar.