Oasnet

Abstract

Influence maximization in a social network is to target a given number of nodes in the network such that the expected number of activated nodes from these nodes is maximized. A social network usually exhibits some degree of modularity. Previous research efforts that made use of this topological property are restricted to random networks with two communities. In this paper, we firstly transform the influence maximization problem in a modular social network to an optimal resource allocation problem in the same network. We assume that the communities of the social network are disconnected. We then propose a recursive relation for finding such an optimal allocation. We prove that finding an optimal allocation in a modular social network is NP-hard and propose a new optimal dynamic programming algorithm to solve the problem. We name our new algorithm OASNET(Optimal Allocation in a Social NETwork). We compare OASNET with equal allocation, proportional allocation, random allocation and selecting top degree nodes without any allocation strategy on both synthetic and real world datasets. Experimental results show that OASNET outperforms these four heuristics.

Overlapping community‐based particle swarm optimization algorithm for influence maximization in social networks

“…For example, Cao et al. [22] proposed a community‐based method called OASNET for solving the problem of influence maximization in social networks. In OASNET, the influence maximization problem was transformed to an optimal resource allocation problem.…”

“…Besides the above greedy or heuristic algorithms, community-based algorithms [22][23][24][25][26][27][28][29][30][31][32][33][34] have also shown their promising performance in solving the social influence maximization problem, which can achieve a good balance between effectiveness and efficiency by utilising the attributes of community structures. These community-based algorithms usually consist of the following three stages: (1) detecting community structures by using a community detection algorithm; (2) utilising the detected community structures to generate candidate set of nodes; (3) selecting seed nodes from the obtained candidate set.…”

“…Note that the importance of overlapping nodes is fully considered in the proposed algorithm OCPSO, the two yellow non-overlapping nodes {1, 11} and one overlapping red node {8} are the final selected seed nodes. Then, in order to measure the quality of these algorithms, we adopt a popular method which calculates the influence spread of the ultimate seed set of each algorithm by running Monte-Carlo simulation for 10,000 times under the F I G U R E 1 (a) Two non-overlapping community structures obtained by a non-overlapping community detection method named LPA [37], where three nodes {1, 2, 11} marked in yellow are the selected seed nodes by one representative community-based algorithm named CGA [22] (suppose k = 3). (b) Two overlapping community structures obtained by an overlapping community detection method named SLPA [38], where two non-overlapping nodes {1, 11} marked in yellow and one overlapping node {8} marked in red are the selected seed nodes obtained by the proposed community-based algorithm overlapping community-based particle swarm optimization (OCPSO) where the importance of overlapping nodes is fully considered.…”

“…Suppose the number of selected seed nodes is set to 3 (i.e. k = 3), the three yellow nodes {1, 2, 11} are the final selected seed nodes obtained by one representative community‐based algorithm named CGA [22]. Note that CGA uses LPA to get non‐overlapping community structure, which neglects the overlapping information.…”

A hybrid algorithm based on community detection and multi attribute decision making for influence maximization

“…They transformed the influence maximization problem into an optimal resource allocation problem. Also, they assumed that different communities are independent of each other and influence cannot spread across different communities [11] .Then, they proposed a recursive relation to find the influential nodes in social networks. Zhang et al studied the problem of influence maximization on networks with community structure.…”