Research On Community Discovery Of Mobile Social Networks Based On User Preferences And Trust Relations

With the rapid development of mobile Internet technology,mobile social networks have emerged in large numbers and become an important communication platform in people's life and work.The characteristics of mobile social network,such as rapid release of information,various forms and unlimited time and space,fully meet the needs of people to obtain real-time information,communicate and interact at anytime and anywhere,profoundly affect people's social way,and play an irreplaceable important role in people's daily life and work.In mobile social networks,users will spontaneously form virtual communities.Research on community discovery is of great significance for public opinion monitoring,marketing and personalized service recommendation.At present,most of the community discovery algorithms for mobile social networks divide nodes according to the similarity of node attributes,ignoring the influence of link relationships between nodes on community clustering.The existing community discovery methods that integrate node attributes and link relationships are few and still have some deficiencies.For example,in the process of considering similarity of node attributes,whether the preferences of users are similar or not is judged only according to the published content of users,while a large amount of noise information exists in the published content of users,which will greatly reduce the accuracy of preference similarity judgment.At the same time,in the process of considering the link relationship,there are explicit link relationship and implicit link relationship between users.Due to the existence of a variety of attribute information of users themselves,only according to the explicit relationship cannot fully describe the relationship characteristics between users.In order to reasonably analyze the relationship between users,the relationship description of mobile social network users can be realized by defining the user trust degree and its transmission method,so as to improve the quality of community division of mobile social network.This paper studies the problems in the existing community discovery algorithms.According to the characteristics of mobile social networks and considering node attributes and link relationships,a community discovery algorithm for mobile social networks based on user preferences and trust relationships is proposed.Preference and trust are used to reflect attribute similarity and relationship characteristics between nodes,so that the division results can reflect more real social relationships and have higher preference cohesion and structural cohesion.The specific research contents of this paper are as follows:Firstly,the paper introduces the basic theory of mobile social network,trust theory and several common classical community discovery algorithms,and analyzes the advantages and disadvantages of each algorithm in the community division of mobile social network.And the existing community evaluation methods are elaborated as the theoretical basis for the follow-up study of this article.Secondly,a similarity calculation method for mobile social networks is proposed,which integrates user preferences and trust relationships.In this paper,the interactive information of users in mobile social networks(i.e.publishing,commenting and forwarding content)is taken as the main factor to evaluate user preferences,and the similarity formula of user preferences is derived.On the basis of related trust calculation research,a similarity calculation method based on trust relationship between nodes is given considering similar trust and relationship trust generated by similarity and relationship strength between nodes respectively.The total similarity between nodes in the mobile social network is calculated by combining the above two,and this is taken as an important basis for the community discovery algorithm.Finally,a community detection based on user preferences and trust relations is designed based on mobile social networks,and micro-blog networks are selected for experimental analysis.Louvain algorithm is improved according to the total similarity between nodes,and a new community partition algorithm for mobile social networks is proposed.In the experimental process,three community evaluation indexes such as modularity,structural cohesion and preference cohesion are introduced and compared with the classical community discovery algorithm to verify the effectiveness of CDPT algorithm.