Research On The Influence Of Trust Relationship In Social Networks

Social network plays a more and more important role in modern people's life,and it has become a part of daily life for many people who use Facebook,Twitter for social activities.These online social networks provide convenient platforms for information dissemination and marketing campaign,there for influence maximization is one of the problems of marketing in social networks.Most existing research only consider information dissemination,without considering the effectiveness of the transmission probabilities between nodes.However,in the practical application,we need to ensure the effectiveness of the transmission probabilities to achieve the maximal influence.Different from the traditional way,we take into account both the social relationship and user similarity to obtain trust transmission probabilities and get a kind of trust propagation model.We studied following three problems based on the trust propagation model.(1).Most existing influence maximization algorithm requires that the influence function must satisfy the sub-modularity and monotonicity properties,but in our real life,the propagation probability is not always to be constant.Therefore,it is difficult to guarantee these two properties strictly.Moreover,it is time consuming for conventional algorithms since they use a large number of Monte Carlo simulations to evaluate the influence spread of seed nodes.In order to solve the two problems above,we proposed a general algorithm that does not need to satisfy these properties and designed a pruning strategy to improve the efficiency of the algorithm.We further design a propagation path-based algorithm to improve the efficiency of influence estimation.(2).According to the marketing in reality,we proposed the influence maximization of revenue control in social networks.To solve the problem,we consider the cost of each node in trust relationship propagation model and achieve the expected benefits with influence maximization under cost budget constraints.We consider three different cost functions and design a greedy algorithm and make further improve the efficiency of the algorithm by pruning strategy and the loading strategy to solve this problem.(3).We began to study the problem:how to find the organizers through the social network that is beneficial for community development,such event organizers comprise a small group of people who not only have the relevant skills or expertise that are required for an event,but also can achieve maxi,mum effect in a certain area.In order to solve this problem,we designs two kinds of greedy algorithms and designs an approximate algorithm based on the principle of pigeonhole.