Optimizing Natural Language Interface for Relational Database
Darshil Shah1, D Vanusha2

1Darshil Shah, Department of Computer Science, SRM IST, Chennai (Tamil Nadu), India.
2D Vanusha, Department of Computer Science, SRM IST, Chennai (Tamil Nadu), India.

Manuscript received on 18 April 2019 | Revised Manuscript received on 25 April 2019 | Manuscript published on 30 April 2019 | PP: 131-135 | Volume-8 Issue-4, April 2019 | Retrieval Number: D6364048419/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: This paper is research on developing a natural language interface for the relational databases, (RDBMS), which takes in the natural language in the form of english like statements as inputs from the operator or user and generates answers in the form of SQL(Structured Query Language) queries. Firstly, the input in the form of natural language is parsed using a semantic grammar and later, it is translated into a SQL query by the use of NLP (Natural Language Processing). The interface makes use of RNN (Recurrent Neural Networks) to convert the natural language input given by the user to the Structured Query language. The network is trained on large number of datasets from the WikiSQL, which makes sure that the accuracy of the model is good enough for real time use in the industry and also for the personal use. Stanford NLP and selfmade datasets are also used to an extent to enhance the efficiency and optimise the query formation process. Finally, a database management system is used to find the result set with its own specialized optimization and planning techniques. The database is implemented in the JSON format to reduce the query processing time when the database is too large. The database is also organised in a tiered architecture format so that the query execution time is reduced.
Keywords: Natural Language Processing, Recurrent Neural Networks, Reinforcement Learning, Relational Data, Structured Query Language, WikiSQL

Scope of the Article: Natural Language Processing