Font Size: a A A

Restricted Natural Language Query Interface Based On Semantic Dependence Grammar Analysis Model

Posted on:2018-03-10Degree:MasterType:Thesis
Country:ChinaCandidate:T SongFull Text:PDF
GTID:2348330515497587Subject:Graphic communication engineering
Abstract/Summary:PDF Full Text Request
With the development of modern science and technology,the computer has become a symbol of this era,people more and more strongly hope to communicate with the computer,and natural language is an important medium for communication with the computer,making the natural language understanding as a key technologies for computer mastering user Intentions.Among them,the understanding of Chinese query language is one of the core contents of natural language understanding,which is attracting more and more attention.The natural language query interface allows people to communicate and interact with computers in a restricted domain of natural language,but the development of natural language processing technology is still not mature enough to transform unstructured natural language into structured queries Language,there are more difficulties.Due to the complexity and diversity of Chinese statements,the difficulty of understanding natural language is strengthened.In order to increase the usability of the natural language query interface,we restrict the input of the statement,design and implement a natural language query interface,the user can input the natural language to indicate the query requirements.The interface can transform unstructured natural language input into the structure of the language can be understood by the computer,to achieve the understanding of the user's inquiry intention.In this paper,the understanding of natural query language is based on the database as the supporting form,we use the lexical analysis and syntactic analysis based on database semantics for processing query language,a method of Chinese query language based on semantic dependency analysis model and grammar recognition is proposed.The semantic dependency tree is treated as an intermediate language,and then through the dependency division and grammar recognition to achieve the statement of the natural language,and the query language is interpreted as structured information.This article includes corpus design,lexical analysis,syntactic analysis and some related work:(1)The corpus designed in this paper is a knowledge base on the natural language query system.It is expressed in the form of database,in the representation of knowledge and the construction of knowledge base,emphasizing language knowledge,domain knowledge and database knowledge Integration,taking the database semantics as the key to the entire process.(2)The lexical analysis part puts forward the shortest path word segmentation method based on database semantics.When segmenting word,first priority to match the domain-specific knowledge and then use the general knowledge,which reflect the importance of the different words for the query statement.(3)The paper proposes a Chinese language query language analysis method based on database semantic dependency analysis model and grammar recognition,which combines the rule-based and statistical-based understanding methods to improve the accuracy and perfection of the analysis.By using the dependent grammar to analyze the dependencies of the sequence of words,a semantic dependency tree is generated to determine the dependency relationship among the components in the query language unit,and the target phrase and conditional phrase are extracted.Finally,the corresponding query target and condition are generated by grammar recognition.In the paper,we constructs an information query system to verify the feasibility of the natural language query interface implemented in this paper.Experiments show that the method used in this paper has a good analytical result,the system can effectively deal with a variety of common forms of query statements,with good availability.
Keywords/Search Tags:Natural language query interface, corpus, semantic dependency, grammar recognition
PDF Full Text Request
Related items