Effective and Efficient Generation of Candidate Networks for Top-k Keyword Search in Relational Database

Abstract

Todays’ data over web are retrieved from Relational Database Management Systems (RDBMS). Web retrieval and content analysis by keyword search in relational database has become the preferential way to search in todays’ era of Big Data. Any keyword search system has two main parts including candidate network generation and evaluation. Most research’s attention is on providing an efficient candidate network evaluation algorithm and relies on non-efficient candidate network generation paradigm. In this work, a novel probabilistic candidate network generation algorithm is proposed. From the outcome of our investigation on benchmark databases including new ISC database, it is possible to conclude that the proposed candidate network generation approach enhances the efficiency of databases keyword search.