| Title | A Hybrid Algorithm for Influence Maximization of Social Networks |
| Publication Type | Conference Paper |
| Year of Publication | 2019 |
| Authors | Lin, Yongze, Zhang, Xinyuan, Xia, Liting, Ren, Yue, Li, Weimin |
| Conference Name | 2019 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech) |
| Keywords | activation effect, Big Data, computational complexity, dissemination effect, dissemination process, Encyclopedias, Gnutella node-to-node file sharing network, greedy algorithm, greedy algorithms, Heuristic algorithms, hill climbing, Hybrid algorithm, hybrid distribution value accumulation algorithm, influence maximization, NP problem, optimal solution, optimisation, Peer-to-peer computing, pubcrawl, resilience, Resiliency, Secure File Sharing, social network, social networking (online), social networks, Time complexity, voting situation, wikipedia |
| Abstract | Influence Maximization is an important research content in the dissemination process of information and behavior in social networks. Because Hill Climbing and Greedy Algorithm have good dissemination effect on this topic, researchers have used it to solve this NP problem for a long time. These algorithms only consider the number of active nodes in each round, ignoring the characteristic that the influence will be accumulated, so its effect is still far from the optimal solution. Also, the time complexity of these algorithms is considerable. Aiming at the problem of Influence Maximization, this paper improves the traditional Hill Climbing and Greedy Algorithm. We propose a Hybrid Distribution Value Accumulation Algorithm for Influence Maximization, which has better activation effect than Hill Climbing and Greedy Algorithm. In the first stage of the algorithm, the region is numerically accumulating rapidly and is easy to activate through value-greed. Experiments are conducted on two data sets: the voting situation on Wikipedia and the transmission situation of Gnutella node-to-node file sharing network. Experimental results verify the efficiency of our methods. |
| DOI | 10.1109/DASC/PiCom/CBDCom/CyberSciTech.2019.00087 |
| Citation Key | lin_hybrid_2019 |
A Hybrid Algorithm for Influence Maximization of Social Networks