Multilingual text categorisation for global knowledge discovery using fuzzy techniques

Enabling navigation among a network of inter-related concepts associating conceptually relevant multilingual documents constitutes the fundamental support to global knowledge discovery. To achieve this, a concept-based approach towards multilingual text categorisation using fuzzy techniques is proposed. First, a language-independent concept space for representing documents regardless of languages is generated using a fuzzy multilingual keyword clustering algorithm based on fuzzy c-means. Second, a fuzzy multilingual text classifier that categorises multilingual documents by concepts is developed using the fuzzy k-nearest neighbour classification technique. Referring to the multilingual text categorisation result as a Yahoo-like directory, concept navigation among the whole multilingual document collection is facilitated.