Research On Target Community Discovery In Complex Networks

Target community detection is a personalized and targeted community detection method,which often explores local communities that meet specific conditions according to the sample nodes provided by users.Therefore,it is only focus on communities satisfying the specified conditions,and has a wide range of application scenarios in the real world,such as providing corresponding solutions for personalized recommendation,Internet information supervision,public opinion early warning,etc.The research of target community detection mainly focuses on the prior information extension provided by users,user preferences mining,and target community model construction based on user preferences.Through the summary and analysis of the existing target community detection work,this thesis has achieved the following research results:Firstly,the Target Community Detection with User Preferences and Influence(TCPI)is proposed to locate the most influential and high-quality community related to user's preference.Firstly,the node structure and attribute information are synthesized,and the maximum k-clique containing sample nodes is investigated as the core of the potential target community,and an entropy weighted attribute weight calculation method is designed to capture user preference.Secondly,the inter consistency and the external separability of the community is defined as the community quality function and the highquality potential target community is expanded with the maximum k-clique as the core.Finally,the external impact score of the community is defined,and all potential target communities are ranked according to the quality function and the external impact score of the community,and the communities with higher comprehensive quality are decided as the target communities.In addition,a two-level pruning strategy are designed to improve the performance and efficiency of the algorithm after calculating all maximal kclique.Secondly,we propose a framework named Adaptive Two-level Weighted Target Communities Detection with Double View(ATC-DV)for attributed graph,which can incorporate user preference into community detection,thus steering the algorithm to detect more interesting and densely connected communities with high attribute semantic similarity.The algorithm consists of three phases:(1)Augmentation of user given prior information: the candidate node path containing a finite sequence of nodes similar with user-provided exemplar nodes is established to make up a network backbone;(2)User preferences learning and view weighting: an adaptive two-level variable weighting clustering mechanism is designed for the network backbone to simultaneously compute weights for both view and variables;and(3)Refinement of the target community model: the quality score of target communities is defined based on the combination of inter consistency and external separability.Extensive experiments are conducted to show the effectiveness and applied value of proposed method.