TY - JOUR AU - Appel, Ana Paula AU - Paterlini, Adriano Arantes AU - Traina Jr., Caetano PY - 2010/10/06 Y2 - 2024/03/28 TI - RDBMS as an Efficient Tool to Mine Cliques on Complex Networks JF - Journal of Information and Data Management JA - JIDM VL - 1 IS - 3 SE - Regular Papers DO - 10.5753/jidm.2010.1282 UR - https://sol.sbc.org.br/journals/index.php/jidm/article/view/1282 SP - 407 AB - <br />Complex networks are intrinsically present in a wide range of applications. <br />Real world networks have several unique properties, such as, sparsity, node degree distribution, which follow a power law and a large amount of triangles that further form larger cliques.<br />Triangles and cluster coefficient, which are usually used to find groups, are not always enough to distinguish a different node neighborhood topology.<br />By using cliques of sizes 4 and 5, it is possible to study how triangles become involved to form large cliques.<br />To retrieve these cliques called k4 and k5 a novel technique called ``FCR - Fast Clique Retrieval'' has been developed, taking advantage of the data management and optimization techniques of a relational database management system and SQL to query cliques of sizes 4 and 5. This paper demonstrates that cliques (3, 4 and 5) follow interesting power laws that allow identifying nodes with suspicious behaviors. It also presents an extension of the cluster coefficient formula, which may become a valuable equation to identify nodes that most influence the network first eigenvalue. ER -