Research On Bilingual Dependency Relationship Mapping For Chinese-english Lexicon Construction

As a valuable resource for language processing, bilingual lexicons are widely usedin the area of Natural Language Processing, such as cross-language information retrievaland machine translation, etc. Also, the extraction of bilingual lexicons has become amajor issue in the field of linguistics. Based on the study of the construction of bilinguallexicons in comparable corpus, the paper puts forward a method to construct theChinese-English lexicon based on the maping of contextual dependency relationship, andfurther demonstrates the superiority of the model of bilingual dependency relationship bycomparison. The main contents of this paper include:1) Implement the construction of Chinese-English lexicon based on the traditionalcontext model. This is done by extracting the contextual words of a target word in thesentence.2) Implement the construction of Chinese-English lexicon based on the dependencycontext model. The benchmark system is implemented by extracting the previous andnext nodes of the target word from the dependency tree structure and taking them as thecontext information. Also, the influence of window size and the computational method ofsimilarity is investigated.3) Implement the construction of the Chinese-English lexicon based on the mappingof dependency relationship. Based on the benchmark system, the methods of building themappings of dependency relationships manually and automatically are proposedrespectively, and the construction of bilingual lexicon is implemented by matching thecontextual words and dependency relationships at the same time.By comparing the construction methods of Chinese-English lexicon, this paperconcludes that the performance of the Chinese-English lexicon constructed based on themodel of dependence relationship mapping is significantly improved.