Natural language querying over databases using cascaded CRFs

Abstract

Retrieving information from relational databases using a natural language query is a challenging task. Usually, the natural language query is transformed into its approximate SQL or formal languages. However, this requires knowledge about database structures, semantic relationships, natural language constructs and also handling ambiguities due to overlapping column names and column values. We present a machine learning based natural language search system to query databases without any knowledge of Structure Query Language (SQL) for underlying database. The proposed system - Cascaded Conditional Random Field is an extension to Conditional Random Fields, an undirected graph model. Unlike traditional Conditional Random Field models, we offer efficient labelling schemes to realize enhanced quality of search results. The system uses text indexing techniques as well as database constraint relationships to identify hidden semantic relationships present in the data. The presented system is implemented and evaluated on two real-life datasets. © 2010 Springer-Verlag.