Research On Mining Structural Hole Spanners In The Weighted Networks And Its System Implementation

For the past few years,following fast development of various kinds of networks and computer science,the network scale shows the exponential growth.The idea of what kind of position taken up in the network could obtain the profit has been concerned by many people.The intermediary among individuals or organizations may have access to abundant information and get command of their network tie in order to become a main part of the center in the network so as to acquire significant benefits.As a key concept of network structure and analysis,those structural hole spanners play important roles in acquiring effective network information.Once the structural hole theory is proposed,it is drawn researchers,attentions and concerns in sociology,psychology and economics fields and widely applied in various sectors.The nodes at the structural hole position can obtain cumulative and non-repetitive information,so it is very meaningful to find and use the structural holes spanners.At present,researches on mining structural hole spanners by investigators are particular emphasis on unweighted network rather than weighted network.In an actual network,the edge weight has a great influence on network performance.Studies on edge weight can more actually reflect the characters of real network structure.Therefore,the research emphasis of this dissertation is mainly reflected in two aspects.Firstly,based on weighted network as an object of study,the edge weight is introduced into the network in order to put forward a weighted betweenness centrality based structural hole spanner mining method to discover those structural hole spanners in the weighted network.This method first works out the weighted network community by using the community to discover the algorithm and then introduces the weighted indexes for unweighted indexes of metric structural hole spanners,which will be deemed as important nodes of initialized community to perform the constantly incremental mining structural hole spanners via significance of the node and possibility of the node mining the structural hole spanners.Finally,the effectiveness of algorithm proposed in this dissertation is validated by public data and instance data.Secondly,a system of mining structural hole spanners is designed and implemented in this dissertation.The system has not only realized W_BCHIS algorithm proposed in this dissertation,but also integrated HIS and MaxD algorithm in two unweighted networks,as well as W_HIS algorithm in a weighted network.It has also provided the figure comparison function for experimental results.On account of B/S framework,the system has adopted Java development language.The major tasks of this dissertation are shown as follows:1.To introduce the mining theoretical knowledge and relevant algorithm of structural hole spanners.It mainly introduces structural hole spanners related mining technology and theoretical background knowledge,including measure indexes of structural hole and algorithm of mining structural hole spanners.2.To come up with an acquisition method of structural hole spanners(W_BCHIS)based on weighted betweenness centrality specific to weighted network studies.Weight value is introduced into the classical betweenness centrality in order to make the betweenness centrality algorithm apply to the weighted networks.This method has initialized the node’s importance of each community using weighted betweenness central value;then discovered the structural hole spanners using constant iteration of the node’s important functions and possible functions of node’s occupying structual hole,and performed the experiment using public data set and instance data set.The results show that the algorithm proposed in this dissertation is able to more accurately discover those structural hole spanners in the network.3.To design and implement the system of mining structural hole spanners.The major functional modules of the system consist of mining structural hole spanner module,community detection module and visualization module.Based on B/S framework,the major development language of the system is Java.The main function of the system is achieved by using MVC design mode.Taking mining structural hole spanners as an example,a few key links in implementation have been set forth in details,such as files upload,code realization of algorithm in mining structural hole spanners and dynamic diagram generation,etc..In the end,the system is tested using cases to display the page effects after system implementation.