Research On The Construction Of A Nlp-oriented Chinese Sentence Semantic Knowledge Database

The research on the construction of NLP-oriented Chinese sentence semantic knowledge database can promote the development of natural language processing (NLP), as well as deepen the semantic research on modern Chinese.Based on analyzing and concluding the existing sentence semantic resources at home and abroad, this paper raises a NLP-oriented Chinese sentence semantic knowledge system on principles of keeping up with the development of natural language processing, being application-oriented, and conforming to the characteristics of Chinese, and tries to meet the needs of Chinese information processing. And then on the basis of this sentence semantic knowledge system, the construction of a NLP-oriented Chinese sentence semantic knowledge database is explored theoretically and practically.The article is divided into five parts altogether, and the main contents of each part are arranged as followed:Chapter one, introduction, introduces five important sentence semantic projects at home and abroad from the aspects of target, method, information tagging, composition, application or influence. Then, some problems existing in the construction of Chinese sentence semantic resources are concluded. The method and value of this research are also presented in this part.Chapter two, three and four are the main parts of this paper.In chapter two, the sentence semantic knowledge needed by NLP, first the NLP's demands to language knowledge are analyzed, and based on this a NLP-oriented Chinese sentence semantic knowledge system is constructed. It consists of two subsystems, the system of semantic knowledge within sentences and the system of semantic knowledge between sentences. Also the main elements of these two subsystems are introduced.Chapter three discusses the inner-sentence semantic knowledge database. First, analyzes semantic roles, which is the core element of the inner-sentence semantic knowledge database by its fineness hierarchy, number setting and classification, and put forward an evaluation criterion for semantic role system. Then, the construction of the Chinese syntactic-semantic tagged corpus is introduced, including its target, principle, method, procedure, source of corpus and information tagged.Chapter four aims at the inter-sentence semantic knowledge database. In this part, the article first defines the term textual entailment and its types based on the introduction of the research status abroad, then introduces the target, method, procedure, source of corpus, information tagged and quality assurance in detail, and makes a statistical research based on the Chinese textual entailment corpus.Chapter five is the conclusion, where the summary of this paper is made, including the research result, its inadequacies and next research tasks.The innovation of the thesis lies in: With an eye to the needs of NLP, this research tries to construct the Chinese sentence semantic knowledge database from two different angles, inner-sentence and inter-sentence. As a result of the theoretical researches and practical explorations, it will supply Chinese information processing with an available knowledge system and accumulate experience for the construction of large-scale sentence semantic knowledge databases in future.