Design of an ACO algorithm for Solving Community Finding Problems

Abstract

The amount of data generated by social media users is increasing exponentially mainly produced by the high number of users connected everyday that interacts with each other through the Social Network (SN). As a result, SNs has become an interesting domain for research due to the wide variety of problems that can be solved. Among these problems, this work is focused on Community Finding Problems (CFP) whose goal is to group the different users in several clusters in such a way users belonging to the same cluster are similar (according to a specific metric) whereas they are different from the users of the other clusters. In this work, we describe the algorithm proposed in [1]. This algorithm for CFP is based on Ant Colony Optimization (ACO) algorithm, and it uses the information regarding the topology of the network, i.e. the connections of the users in the SN. For the experimental phase, we have compared the performance of the described algorithm against the performance of some well-known algorithms extracted from the tate-of-theArt. The results reveal that the proposed Topology-based ACO algorithm is a good approach to solve the community finding problem and it provides competitive results against the analyzed algorithms

Antonio Gonzalez-Pardo

Associate Professor

Lecturer at the Computer Science Department. Main research interests are related to Computational Intelligence and Metaheuristics applied to Social Networks Analysis, and the optimization of graph-based problems.