NLKBIDB - Natural language and keyword based interface to database

One of the major source of information is database which plays an important role in computer field. Almost all IT applications are storing and retrieving information from databases. There are various interfaces available to retrieve data such as form based, natural language and keyword based. Data retrieval from the database requires knowledge of database language like SQL [1]. The need of natural language arises from the fact that common people (other than SQL experts) are not used to writing SQL query. Therefore researchers have identified a new system called Natural Language Interface to Database (NLIDB) and Keyword Based Interface to Database (KBIDB). In this paper we have proposed architecture of a Natural Language and Keyword Based Interface for Database (NLKBIDB) which provides solution for syntactically correct and incorrect natural language input query. Our partial experiment of Lexical Analyzer and Keyword based interface on agriculture survey database solves 53% of syntactically incorrect query which will not be solved by natural language interface resulting in increase of rate of SQL query conversion.

I. Introduction

People around the world need to deal with database for data retrieval. There are diverse solutions of user interface to retrieve the data from database i.e. Form based interface, Natural language based query interface, Keyword based query interface. Form based interface provides the interactive user interface but it gets failed if provider has not provided retrieval of data variations which is possibly manipulated by SQL. Natural language or Keyword based interface allow the user to access data by entering query in natural language either in English or in any other language. Only experts can know how to retrieve the data using SQL query command but what about the person other than experts who want to retrieve the data from the database without the knowledge of SQL. To facilitate the interaction between the non-professional and database, Natural language becomes necessity. Main objective of NLIDB is to accept the query sentence and try to understand it by applying lexicon, syntactic and semantic analysis and then converts into SQL [1]. Keyword based interface accepts query type of search engine query which retrieves keyword from the input query and converts into SQL by applying rules from the generated knowledge base. Natural Language Interface is applied on various formats of data for natural language input query i.e. XML [2], Database Management System (DMBS) [3], Relational Database Management System (RDBMS) [4], [5], [6], [7], [8], and Object Oriented Database [9] but majorly past developed Natural Language Interface is using RDBMS because 95% of the database based applications used in recent times are based on relational models referred as relational database [10].