【摘要】 为提高社区发现算法识别中心节点的精度及算法鲁棒性，提出一种基于有效度局部协同优化的社区发现算法(LCED)。设计有效度指标，通过多次局部扩展使社区结构自然划分出来。在局部扩展过程中，给出节点对的吸引力函数作为节点访问顺序的依据，提高算法鲁棒性；改进适应度函数解决过度融合问题，分配冲突节点优化社区结构。算法在真实网络和合成网络上的对比实验结果表明，LCED算法具有更高的准确率和鲁棒性。更多还原

【Abstract】 To improve the accuracy of the community detection algorithm to identify the center nodes and the robustness of the algorithm, a local collaborative optimization based on the effective degree for community detection algorithm(LCED) was proposed. The effective degree index was designed, and the community structure was naturally divided by multiple local expansions. In the process of local expansion, the attraction function of node pairs was given as the basis of node access order, which improved the robustness of the algorithm, and the adaptation function was improved to solve the over-fusion problem. The conflict nodes were assigned to optimize the community structure. Experimental results of comparing algorithms on real and synthetic networks show that the LCED algorithm has higher accuracy and robustness.更多还原