摘要
复杂网络中重要节点的发现在实际应用中有着重要意义。重要节点的发现依赖于节点的重要性评估,而如今的一些节点重要性评估参数如介数、度分布、聚类系数等存在适用范围有限、评估结果不够全面等局限性,因为节点在复杂网络中的重要性程度不仅仅只受单一因素的影响。提出一种基于拉普拉斯特征映射算法的节点重要性综合评估方法。这种方法能综合考虑全部节点的局部特征,可以准确地对节点在复杂网络中的重要性程度进行评估,同时能够得到良好的结果。由于该方法无需对复杂网络中所有节点的全部特征进行评估,极大地缩减计算时间,在保证精确性的同时,提高效率,并通过MATLAB仿真与现有算法的结果进行分析对比,证明该算法的有效性和可行性。
In complex networks, it is important how to evaluate the nodes according to their importance. Discovering the important nodes depends on the importance evaluation of nodes. Nowadays most of the existing methods of evaluating pivotal nodes （e.g. betweenness-based, degree distribution, clustering efficient） only think about one factor but not the integration of all complex network in evaluating the importance of nodes, so this methods each can do work in limited application scope. Proposes Laplacian Eigenmaps algorithm to discover the vital nodes in complex networks. In this algorithm, ranking key nodes will consider each node＇s integration of whole complex network. After that, it can be accurate to evaluate important degree of the node in the network, and can get a good result. Due to the evaluation does not need calculate whole attribute of complex networks＂ nodes, so it can greatly reduce the computing time, and the efficiency to discover the key nodes can be improved. Finally, the experiment＇s result of MATLAB simulation about the proposed algorithm shows that this algorithm is feasible and effective.
出处
《现代计算机》
2017年第6期7-10,共4页
Modern Computer
