Method for establishing link between natural language question and database

By obtaining user input problems, database and sample data sets, using vectorization and similarity calculation to determine the reference data and candidate tables, and recalling candidate fields through field recall model, the problem of low accuracy of natural language problems and database links is solved, and more accurate SQL statement generation and more efficient data access is achieved.

CN119719133APending Publication Date: 2025-03-28BEIJING KNOWLEDGE ATLAS TECHNOLOGY CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202411771880.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-04
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

In the prior art, when establishing links between natural language problems and databases, there is a problem of low accuracy, mainly due to uncertainties such as word meaning ambiguity, syntactic ambiguity and insufficient information in natural language.

Method used

By obtaining user input problems, database and sample data sets, using vectorization and similarity calculations to determine the reference data and candidate tables, and recalling candidate fields through field recall model to establish a link between natural language problems and database.

Benefits of technology

It improves the link accuracy between natural language problems and database tables, provides a more accurate basis for generating SQL statements, and improves the convenience of data access and user experience.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to a method for establishing a link between a natural language question and a database, and belongs to the technical field of natural language process.The method comprises the steps that an input question of a user, the database and a sample data set are obtained, and sample data in the sample data set comprises the natural language question and a corresponding SQL statement; determining reference data according to the input question and a natural language question in the sample data; selecting a first candidate table according to SQL statements in the reference data; selecting a second candidate table according to the input question and a table name in a database; determining a field recall model according to the quantity of the sample data; and inputting the input question, the first candidate table and the second candidate table into the field recall model to obtain candidate fields. According to the method and the device, the link between the input natural language question and the table in the database is accurately established, so that a basis is provided for accurately generating the SQL statement.
Need to check novelty before this filing date? Find Prior Art