Ontology-based natural language interface to public security population database
Xiangwu Ding, Hao Liu · Atlantis Highlights in Computer Sciences/Atlantis highlights in computer sciences · 2023
Natural Language Interface to Database (NLIDB) could convert natural language queries into SQL automatically, which has been extensively studied.However, how to apply NLIDB to the public security population database (PSP-DB) remains an open problem due to the challenges to utilize domain knowledge and generate complex queries involving multiple tables.To tackle these problems, this paper proposes an ontology-based NLIDB approach combining with public security population ontology (PSP-Ontology) and syntactic analysis.Its key idea includes: (1) constructing an ontology from the schema of PSP-DB and extending it with synonym expansion; and (2) proposing an association path processing algorithm to handle multi-table connection path in SQL generation.We have evaluated the approach on the population database from Shanghai Public Security Bureau.The results show that the PSP-Ontology and association path processing algorithms could alleviate these two challenges and improve the accuracy of SQL translation effectively.