CAJ | 학술논문

为提高单机处理复杂网络规模的能力，提出一种新的重叠社区发现算法。首先，通过基于图压缩的社区结构表示模型(压缩社区图)，对网络进行无损压缩；然后，在压缩社区图上基于种子迭代的思想，通过不断优化社区适应度函数将种子扩展成社区；最后，将相似度高的社区进行合并，得到最终的重叠社区结果。由于压缩后的凝聚图大大降低了待处理的网络规模，并能在一定程度上减少重复计算，该方法可以大大提高计算效率和单机处理的网络规模。
위제고단궤처리복잡망락규모적능력，제출일충신적중첩사구발현산법。수선，통과기우도압축적사구결구표시모형(압축사구도)，대망락진행무손압축；연후，재압축사구도상기우충자질대적사상，통과불단우화사구괄응도함수장충자확전성사구；최후，장상사도고적사구진행합병，득도최종적중첩사구결과。유우압축후적응취도대대강저료대처리적망락규모，병능재일정정도상감소중복계산，해방법가이대대제고계산효솔화단궤처리적망락규모。
To improve the capacity of single machine to handle complex netWork,overlapping communities detection algorithm Was proposed.First,a graph compression based social netWork model,namely agglomerative graph,Was introduced,Which Was a lossless compression to the original netWork.Then,inspired by the idea of iteration based on seeds,the selected seeds Were expanded to the communities by optimizing the proposed community fitness function iteratively. Finally,the communities of high similarity With each other Were merged to get the final results. Since the scale of the netWork to be dealt is significantly reduced,and some redundant computations are avoided,the proposed algorithm is of high efficiency.