Keyword Search in Graphs: Finding r-cliques

Keyword search over a graph finds a substructure of thegraph containing all or some of the input keywords. Mostof previous methods in this area find connected minimaltrees that cover all the query keywords. Recently, it hasbeen shown that finding subgraphs rather than trees canbe more useful and informative for the users. However, thecurrent tree or graph based methods may produce answersin which some content nodes (i.e., nodes that contain in-put keywords) are not very close to each other. In addition,when searching for answers, these methods may explore thewhole graph rather than only the content nodes. This maylead to poor performance in execution time. To address theabove problems, we propose the problem of finding r-cliquesin graphs. An r-clique is a group of content nodes that coverall the input keywords and the distance between each twonodes is less than or equal to r. An exact algorithm is pro-posed that finds all r-cliques in the input graph. In addition,an approximation algorithm that produces r-cliques with 2-approximation in polynomial delay is proposed. Extensiveperformance studies using two large real data sets confirmthe efficiency and accuracy of finding r-cliques in graphs.