Research On Social Network Marketing Strategies Based On Algorithms

Social network is one of the most popular applications on the Internet,which includes almost all kinds of user-centric network service forms.With the popularization of various kinds of social network applications,social network draws the attention of various advertisers because of its remarkable advantages,such as abundant users,fast information spread,low cost,great impact and strong interaction.Advertisers in social network marketing can effectively improve their gains by using some methods,including personalized recommendation,influence maximization and game theory.According to the different demands of advertisers,this thesis uses collaborative filtering algorithm,influence maximization algorithm and game theory to respectively propose four kinds of marketing strategies:personalized recommendation,topic-concerned influence maximization,multi-social networks influence maximization and optimal investment in co-marketing.Firstly,the thesis proposes a personalized collaborative filtering recommendation strategy based on user interest profiles.This strategy makes full use of user data in social network to build social user interest profiles and further proposes a collaborative filtering algorithm based on social user interest profiles by combining the similarity between social user interest profiles into the collaborative filtering algorithm.The experimental results indicate that the proposed strategy achieves a higher accuracy of recommendation.Secondly,the thesis proposes a topic-concerned influence maximization strategy by exploring information diffusion model IC-ITUI.We firstly propose a novel information diffusion model IC-ITUI and solve specific topic information maximization problem in the proposed model.The experimental results indicate that the proposed strategy can help the advertiser to find marketing initial seed nodes and then maximize the spread of their marketing information.Thirdly,the thesis establishes a multi-social networks information influence maximization strategy based on aggregate graph.This strategy aggregates multiple social networks into a network graph according to bridge users and then solves the problem of influence maximization on multiple social networks based on aggregate graph.The experimental results show that the proposed strategy can find marketing initial seed nodes and then maximize the spread of marketing information on multi-social networks.Finally,the thesis provides an optimal investment strategy in co-marketing based on Stackelberg game for the advertiser.This strategy builds one-leader multi-followers Stackelberg game model to model the relationship between advertiser and social network platforms and further proposes an algorithm based on egalitarian welfare solution to achieve Stackelberg equilibrium of proposed game model,which is the optimal strategies for both advertiser and social network platforms in the process of co-marketing.Furthermore,the proposed game model can maximize the advertiser's profit in its co-marketing with multiple social network platforms.The numerical analysis shows that the advertiser can obtain higher utility in this strategy compared with other models.