Ranking of keyword‐combined searches in relational databases based on relevance to the user query

Abstract

In this Letter, the authors deal with ranking of keyword-combined searches in relational databases based on relevance to the user query, which they call KEYSIM searches. They formally define KEYSIM searches and propose a threshold-based method for efficiently processing KEYSIM searches. Their proposed method is the first one to find top-k results considering both numerical similarity and textual similarity. Through the experiments using five real and synthetic data sets, they show the efficiency and scalability of the proposed method.