Research On The Query Information In The Chinese Query Sentences Of Database

Natural language interface to database systems (NLIDB) is one of the earliest and broadest fields that have been researched in natural language process (NLP). It is the field of great theoretical and practical value that mainly involves the disposal of natural language and the research of database system.In the research of the query interface of NLIDB, the comprehension to the query sentences entered by the user is the key part. And the understanding of the query condition and target are the main sections of this key part. The aim of this paper is to investigate the recognition of the query target and condition of the query sentences, which includes:One is the systematic study on the implicit knowledge of the database. Corresponding index to the knowledge is also proposed. Database implicit knowledge refers to the knowledge that comes up because of the relevance among the data and can be represented with natural language. This part of data exists in the database but the implicit knowledge can only be got through certain operations or logical calculations. The implicit knowledge of the database reflects the comprehension of the user to the real world. Consequently, the user will resort great amount of the implicit knowledge of the database in the course of query. The ability to deal with the implicit knowledge of the database reflects whether it is easy to use. This paper proposes how to improve the ability of processing the implicit knowledge of the database of the query sentences.In the recognition of the query target, the processing ability of the implicit knowledge of the database is improved. The query targets are classified and relevant changing ways of calculation are also proposed. The query target recognition mode of the query sentences based on semantics of the database is presented. This mode can recognize different query targets in different sentence structures.In the recognition of the query condition, the method of recognizing the query condition that centers on the field value and is based on the semantics of the database is proposed. The processing of the query condition that contains implicit knowledge of the database is improved. Detailed research is on the difficult points of the recognizing the query condition, which are recognizing the query condition items that contain domain verbs and that contain negative words. Solution and the method of calculation are presented. At last, the general method of calculating the recognition of the query condition is proposed.