Research And System Implementation Of Influence Source Localization Algorithm Based On Social Network

As a suitable information diffusion platform,online social network plays an important role in the dissemination of information.The spread of positive information on social networks has a great impact on people’s lives.However,the social network itself is open and synthetic,and all kinds of bad,false information and reactionary remarks can be spread and spread across regions and borders,seriously endangering social stability and national security.The problem of localization of influence sources is an important problem in social network research,which aims to find the most influential users who initially disseminate information in the network where influence is propagated.In general,in social networks,various types of diffusion dynamics can be viewed as diffusion processes,and determining the location of diffusion sources is crucial.How to locate the users who initially disseminated influence in the process of information dissemination,and solve the spread of negative influence from the root,this is the problem of influence source positioning.This problem is highly theoretical and complex,and has a wide range of applications,such as epidemic monitoring and public opinion monitoring.Therefore,in order to facilitate relevant departments to obtain and monitor,analyze and locate influence source information in a timely manner in social networks,this paper studies the problem of locating influence sources in social networks from two perspectives.A source localization algorithm based on independent paths is proposed.Secondly,considering the propagation time factor,an influence source localization algorithm based on Steiner tree is proposed,and a corresponding public opinion monitoring and influence source localization analysis system is developed according to the needs of practical application scenarios.The main work of this paper is as follows:(1)A source localization method based on independent path analysis under the IC(Independent Cascade)model is proposed.The method first uses independent paths to calculate the propagation probability between nodes,then establishes the likelihood function of the influence range,and finally uses the greedy algorithm to select the node with the largest likelihood function to join the seed set.By conducting experiments on datasets of real networks and synthetic networks,the results show that our proposed method can achieve more accurate results compared with classical methods and similar methods.(2)A source localization algorithm based on Steiner tree under IC model is proposed.The method first calculates the distances from all vertices to the observation points,then estimates the coincidence between the distance and the activation time,then decomposes the observation node set to obtain the candidate seed set,and finally uses the Steiner tree structure and the idea of k-means clustering to Optimize the candidate seed set.Through experiments on datasets of real networks and synthetic networks,the results show that our proposed method outperforms other methods in every evaluation metric.(3)Based on the algorithm proposed above,this paper designs and implements an influence source localization system in social networks.The system uses front-end development technology to realize different functions such as information source simulation,diffusion and source location.