EFSLIM: INFLUENCE MAXIMIZATION USING ENHANCED SHUFFLED FROG LEAPING APPROACH IN SOCIAL NETWORKS

초록

Social networks are influencing people to make choices and decisions on others. With the advertisement of the business products on the web and other sources, the development of social network has increased tremendously. Many social networking organizations develop their network nodes by using a popular concept known as Influence Maximization, which is a greedy approach. The objective of this approach is to maximize the nodes by identifying minimum subset nodes formed at the base level, which has the capability to influence other nodes. The existing algorithm, Independent Cascade Model, in which the activation probability of every node is computed and an influential set is generated based on the behaviour of other nodes due to the influence of the parent nodes. The major disadvantage of this mechanism is the potential creation of vulnerable nodes which spread the information without knowing the adverse effect on the individual. For example, advertising the junk food attractively may have impact on the obese person. The issue with this approach is influencing the entire population using vulnerable nodes. The proposed model tries to influence the targeted audience by maximizing the non vulnerable nodes in the graph. Since the interaction is associated with the behavioural patterns of the individuals, the model uses the genetic algorithm termed as Enhanced Shuffled Frog-Leaping. It searches the local space by encrypting the cumulative responses from other nodes and it updates the fitness function based on the utility. It is evident from the obtained experimental results that the proposed Enhanced Shuffled Frog Leaping Approach for Influence Maximization (EFSLIM) in social network showed the influence spread and statistical tests as an effective and advanced model for overcoming the influence maximization problems. The proposed model showed better performance and reached 1400 of spreading size for 100th node but the existing DFLA obtained 800 of spreading size. © 2022 Little Lion Scientific.

키워드