CAJ | 학술논문

基于密度吸引点和其对相邻节点的影响度，提出了一种密度分布社区发现算法。该算法以节点度数最大的密度吸引点为初始社区，访问社区的相邻节点，把对社区影响度最大的节点加入到社区中，如果有些节点对多个社区都有影响，则把它归属为影响度最大的那个社区中，同时如果两个社区之间的相互影响度很大，可以将这两个社区合并为一个社区。将该算法应用到 Zachary 空手道俱乐部网络和随机无标度网络中，实验表明该算法能够很好地分出网络中的社区，同时实验还发现社区的收敛速度与幂率分布特性近似成反比。
기우밀도흡인점화기대상린절점적영향도，제출료일충밀도분포사구발현산법。해산법이절점도수최대적밀도흡인점위초시사구，방문사구적상린절점，파대사구영향도최대적절점가입도사구중，여과유사절점대다개사구도유영향，칙파타귀속위영향도최대적나개사구중，동시여과량개사구지간적상호영향도흔대，가이장저량개사구합병위일개사구。장해산법응용도 Zachary 공수도구악부망락화수궤무표도망락중，실험표명해산법능구흔호지분출망락중적사구，동시실험환발현사구적수렴속도여멱솔분포특성근사성반비。
Based on the density attractor and its effect on adjacent nodes , this paper proposes a community discovery algorithm of density distribution . The algorithm takes the highest density attractor of the node degree as the initial community , and then visits the adjacent nodes of communities , adding the most influential node to the initial community . If some nodes have influence on sev-eral communities , then it belongs to the most influential community . At the same time , two communities can be merged into one community , if the influence is big enough between these two communities . Applying the algorithm to Zachary karate club network and random scale-free network , the experiments show that this algorithm can divide communities of network; meanwhile , it is also found in the experiences that the rate of convergence speed of the community is inversely proportional to the power distribution characteristics .