International Journal of Engineering and Technical Research (IJETR) ISSN: 2321-0869, Volume-3, Issue-6, June 2015 213 www.erpublication.org Abstract— Many complex systems in the real world can be modeled as signed social networks. Community detection in signed social networks is a challenging research problem aiming at finding groups of entities having positive connections within the same cluster and negative relationships between different clusters. Many community detecting algorithms have been developed in the past. But, most of them are only effective for networks containing only positive relations and, are not suitable for signed social networks. This work is primarily for the networks having both positive and negative relations; these networks are known as signed social network. In this work DFA (detection and formation algorithm) has been proposed which works in two phases. The first phase is based on Breadth First Search algorithm which makes community structure on the basis of the positive links only. The second phase takes the output of first phase as its input and produces community structure on the basis of a robust criteria termed as participation level. Proposed algorithm can find the signed social networks where the negative inter-community links and the positive intra-community links are dense. Proposed algorithm is also useful in detecting the communities from only positive conventional graphs. Moreover it doesn’t require any external parameter for its operation as is the case with other algorithms like FEC (finding and extracting communities). Inclusion of a new node in the graph is tackled effectively to reduce the unnecessary computation. This algorithm proceeds in breadth first way and incrementally extracts communities from the network. This algorithm is simple, fast and can be scaled up easily for large social networks. The effectiveness of this approach has been demonstrated through a set of rigorous experiments involving both benchmark and randomly generated unsigned and signed networks. The algorithm is simulated by using GUESS (Graph Exploration System) tool. Results provided by proposed algorithm are good and comparable with other algorithms for unsigned and signed social networks in terms of accuracy and order of time complexity. Index Terms—Community Detection, Community Structure Social Networks, Signed Social Networks, I. INTRODUCTION Social networks are formed by individuals having some properties in common. A social network can be defined as a graph G = (V, E), where V = (v 1 ,v 2 ,v 3 , . . . v n ) is the set of Himanshu Sharma, Computer Science & Engineering Department, Arya College of Engineering & Information Technology, Jaipur, Rajasthan, India, 9460867761. Vishal Srivastava, Computer Science & Engineering Department, Arya College of Engineering & Information Technology, Jaipur, Rajasthan, India, 9214052387. vertices, and E = (e 1 ,e 2 ,e 3 , … e n ) is the set of edges connecting pairs of vertices. For example, in a human social network, each vertex (node) denotes an individual, and each edge (link) denotes a relation between two nodes. In weighted social networks, each link is attached with a real number called weight which represents in some sense how closely connected the vertices are [7]. In the field of social science, the networks that include only positive links are also called positive social networks, and the networks with both positive and negative links are called signed social networks [3] or signed networks for short. A signed social network in its simplest form can be viewed as a weighted bidirectional graph having three types of weights {+1,0,-1} [3]. Weight “+1” is assigned to the edges connecting positively a pair of nodes, Weight “-1” is assigned to the edges connecting negatively a pair of nodes and Weight “0” is assigned if an edge does not exist between the nodes. For example, a network of nations where positive relation shows the political alliance and negative shows the opposition. In the friends-enemies network, positive link shows that they are friends and the negative shows that they are enemies. In the literature, there are a number of examples of weighted graphs in which the weights assigned to the edges lies in a particular range of numbers. However, these graphs may be considered as a special case of the previously explained signed graph, we can transform these types of graphs to simple signed graphs by assigning +1 to the weights above a predefined threshold and -1 to the weights less than that level. This generalization of social networks is done because normally it is not easy or fair to give weights to the relationship of an individual with other individuals. II. PREVIOUS WORK ON DETECTING COMMUNITY STRUCTURES IN SOCIAL NETWORKS In context of social networks the task of grouping the set of vertices exhibiting similar properties or behavior is referred as Community Detection. Social networks are generally sparse in global yet dense in local. They have vertices in a group structure such that the vertices within the groups have higher density of edges while vertices among groups have lower density of edges. This kind of structure is called the community which is an important network property and can reveal many hidden features of the given networks. Two vertices having the same attribute have a positive link between them and the vertices having the opposite attribute will have a negative link. These vertices are classified on the basis of both the link density and the signs of the link. This task becomes challenging when there are some negative Detecting Community Structures in Signed Social Networks (An Automated Approach) Himanshu Sharma, Vishal Srivastava