Research on interpretability recognition of Chinese discourse markers based on dependency graph
XIAO Ming1,2， XIAO Yi3
(1.Research Center for Language and Language Education，Central China Normal University， Wuhan 430079，China；2. Informatization Office， Central China Normal University， Wuhan 430079，China；3. School of Information Management， Central China Normal University，Wuhan 430079， China)
Abstract：The study of discourse markers in natural spoken language is of great significance for speech interaction， discourse comprehension， sentiment analysis， human-computer dialogue and spoken machine translation. In order to realize the automatic recognition of discourse markers， based on the dependency grammar theory， the semantic and functional information of the syntactic dependency， semantic dependency， chakra position and co-occurrence components of discourse markers are analyzed and determined in this paper. Aiming at the lack of principle and semantic explanatory problems of artificial intelligence deep learning methods， four explanatory machine learning methods of nave Bayes， decision tree， large-scale linear support vector machine and Bayesian network are used to identify and compare discourse markers. The results show that the recognition accuracy of Bayesian network can reach 92.3%， which demonstrates the feasibility and effectiveness of this study.
肖 明,肖 毅. 基于依存关系图的汉语话语标记可解释性识别研究[J]. 华中师范大学学报(自然科学版), 2023, 57(4): 528-538.
XIAO Ming,XIAO Yi. Research on interpretability recognition of Chinese discourse markers based on dependency graph. journal1, 2023, 57(4): 528-538.
2023, Vol. 57 
