Research On Event Arrangement Strategies In Event-based Social Network

Event-based social networks(EBSN)is a new social network centered on events.With the development of big data technology,researchers have used various machine learning and data mining techniques to analyze user behaviors and calculate user interest preferences to implement EBSN event personalization recommendations to solve the problem of “information overload”.However,the existing social network architecture and user needs have undergone tremendous changes.Simple personalized recommendations can no longer meet the needs of existing social networks.Instead,they need a smarter and more organized event scheduling strategy to satisfy social networks.Provide diverse needs.Therefore,the serialization and globalization of the event scheduling problem has become a hot topic of current research.Aiming at the continuous event sequence arrangement,this paper proposes a continuous event sequence arrangement strategy with constraints.That is to arrange a set of reasonable event sequences for users in the case of user permissible time and cost budget.The sequence of sequential events takes into account the degree of user preferences,the category of events and the distribution of events.In addition,the sequence of events,the location of the space,the cost and the conflict have brought great challenges to the solution of the problem.To solve the above problems,we first define the related concepts in the continuous event arrangement problem,then we give the objective function of continuous event arrangement problem,and prove that the problem is a NP hard problem.In order to optimize the target effectively,an approximate algorithm based on greedy strategy is proposed.In order to further solve the problem of data set size,a 0-1 integer programming algorithm based on utility is proposed.Finally,performance tests on two real data sets show that the proposed strategy has a good effect.Aiming at the global event arrangement,this paper proposes a global event arrangement strategy considering constraints.EBSN recommendation systems should take into account the requirements of the users and event organizers at the same time.That is,it should not only meet the users' preference as far as possible,but also ensure global balanced allocation of event resources.Thus,it is very urgent and necessary to study EBSN global recommended strategies.However,due to the multiple constraints in EBSN,including the number of events that users are allowed to participate in,the number of users that events allow to accept and the conflict constraints between events and so on,so the user-oriented traditional recommendation methods are no longer applicable for EBSN global recommendation.In the paper,we first define the optimization object of global recommendation,that is,to minimize the sum of all users' dissatisfaction and all events' dissatisfaction satisfying constraint conditions.Then,in order to transform the constrained problem into a non-constrained problem,we establish a penalty term for each kind of constraints,and get a single objective function.For achieving the optimization object,we suggest a constrained binary particle swarm optimization algorithm(CBPSO),which advantage is that it can solve the binary discrete space problem under multiple constraints conditions.For improving the optimization performance further,an improved binary particle swarm optimization(IBPSO)algorithm and an extended IBPSO algorithm(EX-IBPSO)for global recommendation are proposed.The experiments on a real dataset show that the proposed approaches are available and effective.