Reserch On Social Advertising Algorithm In Online Social Network

With the rapid development of online social networks,the scale of social networks becomes larger and the data becomes increas:ingly richer,which promotes the development of advertisements on social platforms.Advertising has become one of the main profitability of social platforms.How to use social influence to improve the effectiveness of social advertisements,becomes one of the key research issues in online social networking.However,there are many kinds of social advertisements,and existing researches are insufficient in the accuracy,efficiency,and mining social relationships.Therefore,we conduct in-depth research on two types of social advertising algorithms,including social display advertising and viral marketing in this paper.In terms of social display advertising,a topic-based display advertising optimization problem in online social networks is raised for the first time,and a variety of solutions are proposed.Firstly,the topic-based display advertising optimization problem is formally defined and the complexity of this problem is analyzed.Secondly,in order to overcome the shortcomings of the existing heuristic strategy,a topology-based heuristic strategy is proposed,which combines the network topology and topic preference of user.In particular,the topology-based heuristic strategy can be extended to the existing heuristic algorithm.Thirdly,to further improve the efficiency of the algorithm,a community-based solution is designed.Finally,the experiments based on two real datasets are carried out.The experimental results show that the proposed heuristic algorithm improves the expected click rate of advertisements,and the community-based algorithm reduces the running time.In terms of viral marketing,an improved particle swarm optimization algorithm is proposed to solve the location-aware targeted influence maximization,called ILTIM_PSO(Improved LTIM based on Particle Swarm Optimization).Firstly,a forest index is built to find targeted users who are interested in the query topic and region.Secondly,the influencer index is built to select candidate users.Thirdly,the improved particle swarm optimization algorithm is proposed to solve the problem,and the particle related parameters and operations are redefined.The improvements of this algorithm come from three aspects:1)a fitness function suitable for LTIM problem is defined;2)to further accelerate convergence,the set of candidate users is introduced,particles are initialized with probability according to user influence,and particle adjustment strategy based on cross influence is proposed.3)In order to reduce the risk of falling into local optimum and improve the global search ability of the algorithm,the particles are allowed to deteriorate within a certain probability,which refers to the idea of simulated annealing.Fourthly,experiments based on two real datasets are carried out.The experimental results show that ILTIM_PSO has better algorithm accuracy and significantly reduces the running time.