Comparative Analysis of Features in Formal Semantic Conceptual Representation (FSCR) of Modal Verb Based on Causal Attribute Partial Ordered Structure
Wang Liyang, Jianping Yu, Jilin Fu, Tao Zhang · 2024
The progression of natural language processing (NLP) has considerably contributed to the development of general linguistic studies. Among them, the complexity and uncertainty of the semantic system of modal verbs lead to the demand of an effective approach for word sense disambiguation (WSD). In this article, the casual attribute partial order structure (CAPOS) approach is adopted to extract the optimal linguistic features to form formal semantic conceptual representation (FSCR) of English modal verbs. Modal verbs can, may, could, might and be able to, which express the modal meaning of permission, are selected as target words of the comparative study. A WSD rule model with an average accuracy reaching 94.11% is constructed by 1,600 samples sentence randomly selected from a multi-genre corpus of 5 million words. On basis of the features in WSD rules, the FSCRs of modal verbs could be defined and some valuable knowledge about modal semantic class analysis can be mined through the comparative study among modal verbs delivering the same meaning.