| Among all the problems researchers encounter in the area of computational linguistics, lexical ambiguity remains one of the most difficult. It was one of the obstacles that the early machine translation systems foundered upon. It is still one of the problems that stops research from claiming complete success in natural language interpretation.; This dissertation addresses the problem of lexical ambiguity in computer processing of the Chinese language. When the Chinese language is written in Pinyin, polysemy and homonymy are serious problems. A relational semantic model is designed and developed to reduce polysemy and homonymy. The model is based on a semantic network. It adopts the heuristics techniques and makes use of two semantic relations: association and taxonomy. It interacts with the system parser so that syntactic rules, semantic rules and the relation between the two (through parameterization) all function together to achieve disambiguation. |
