Research On Influence Maximization Of Heterogeneous Social Networks Based On Influence Weights And Location Constraints

Influence maximization is an important research direction of social network analysis,which is widely used in advertising marketing,public opinion control and other fields.The influence maximization method aims to find a set of initial seed nodes with high influence,maximize the spread and diffusion of nodes range of influence.At present,the main influence maximization methods are mainly aimed at the homogeneous social network.The homogeneous social network is only a brief description of the relationship between the same objects and objects in the real world,and cannot realy express the relationship between various object types in the real society.social relationship.There are many types of objects in heterogeneous social networks,and various types of relationships between objects contain rich structural and semantic information,which helps to reveal the deep information contained in social networks and promotes the application of maximum influence in practice.At the same time,the complex relationships in heterogeneous social networks bring new chalenges and opportunities to the study of influence maximization.On the one hand,the different relationships between different objects in heterogeneous social networks make it difficult to measure the influence weights between objects,and the influence weights between objects are closely related to the types of objects;on the other hand,applications with location services connect online social networks with real society.Together,the social network is transformed from a virtual world into a complex network of relationships closely related to human life,and the analysis of the relationship between user behavior and geographic location in heterogeneous social networks helps offline merchants to promote their products and services.Based on the above two problems,the main work and innovations of this thesis are summarized as follows:(1)An influence maximization algorithm IMIWH based on influence weights in heterogeneous social networks is proposed.The algorithm proposes the concepts of participation entropy and interaction entropy to measure the influence weight between nodes in a heterogeneous network.Based on this,it measures the influence of nodes in a heterogeneous social network under the linear threshold model.By counting the simple paths from nodes to other nodes in the neighborhood Quantity,which measures the global impact of a node.(2)In a location-based heterogeneous social network,a location-constrained heterogeneous social network influence maximization algorithm TRIM is proposed.First,the TRSAK algorithm is proposed to determine the area formed by the set of locations that are close and highly correlated with the target location;then,the influence of nodes is calculated based on the RR set sampling technique to maximize the influence of the target area.(3)Based on the IMIWH algorithm and the TRIM algorithm,a prototype system for maximizing the influence of the heterogeneous network is realized.First,the social network data set is uploaded and the algorithm parameters are input;secondly,the influence of the nodes in the social network is analyzed through the algorithm module and the seed set is returned;Finally,the seed set is fed back to the system interface to the user.