Chinese Sentence Decomposition Based on Hierarchical Word Order

Abstract

Abstract At present, the field of parsing is occupied by statistical-based methods, and a large number of parsing methods used in English are applied to Chinese. Because of the difference between Chinese and English, it exposes a lot of shortcomings. In dependency parsing, there are many verbs in Chinese sentences, but dependency parsing only uses one core verb, which is not in line with the characteristics of Chinese. Based on the knowledge of Chinese grammar, statistics, and the hierarchical analysis theory of Chinese dichotomy of Zhu Dexi and Lu Jianming, this paper proposes a hierarchical word order-based decomposition of Chinese simple sentences. This method makes up for the deficiency of dependency parsing, and deeply combines the theory of Chinese grammar to decompose Chinese sentences hierarchically in order to extract sentences provides a better basis. Experiments were carried out on the People’s Daily corpus labeled manually by Peking University from January to June 1998. The correct decomposition rate reached 46.692%.