Automatic Concept Mapping between Multilingual Dictionaries and WordNet to Derive Vocabulary for E-Marketplace
Guangyi Xiao, Jingzhi Guo, Zhiguo Gong · 2010
Creating semantically consistent multilingual vocabulary is challenging to concept engineering, natural language process (NLP) and collaborative editing for cross-domain meaning understanding. This paper proposes a novel method of automatic concept mapping (My Lang Mapper) between multilingual dictionaries (same as bilingual dictionary, abbreviated as BD) and the Princeton's WordNet (WN) to derive vocabulary for e-marketplace. It is heuristically developed following the analysis on the senses of words appeared in both WN and BD. By this approach, a word association categorization scheme is introduced to categorize words in WN and BD for achievement analysis, and proposed approach with two patterns of automatic words mapping are created to disambiguate the meaning of words being mapping. This approach's implement and experiment result show that the proposed method is a good complement of the existing collaborative concept editing for vocabulary building method. It suggests that collaborative work for vocabulary building can be significantly reduced.