中文摘要：

由于复杂网络的规模越来越大,在大规模的复杂网络中快速、准确地挖掘出隐藏的社区结构是当前该领域研究的热点问题。目前社区结构挖掘常用的基于快速Newman算法的社区结构挖掘算法之一是一般概率框架方法。以规模日益增大的复杂网络为研究对象,提出了基于GPGPU的一般概率框架并行算法,有效地解决了在大规模的复杂网络中快速、准确地挖掘出隐藏的社区结构问题。实验证明,随着节点数的增加,该并行算法在不损失准确性的前提下运行效率有所提高,为复杂网络社区结构挖掘的研究提供了一种高效的解决方案。

英文摘要：

Due to the increasing scale of the complex network, how to find out the hidden community structure in a large-scale complex network quickly and accurately is becoming the hot topics of current research in this field. One of the community mi- ning algorithms which based on fast Newman algorithm is general probabilistic framework algorithm. This paper focused on the increasing complex network, put forward the parallel computing method of general probability framework algorithm which based on GPGPU, and effectively solved the problem of hidden community structure on the large-scale complex network quickly and accurately. Experiments show that with the increase of the number of nodes, the algorithm not only can runs efficiently under the premise of no loss of accuracy, and can provides an efficient solution for complex network community structure mining.