摘要
确定复杂网络中节点的影响力对于网络上信息传播及网络营销等具有重要的价值。Page Rank算法和LeaderRank算法是两种著名的对复杂网络中节点进行重要性排序的算法。分别使用这两种算法对斯洛伐克最流行的在线社会网络Pokec中的用户进行了重要性排序。与度中心性指标排序结果进行对比,分析了这种排序结果出现的原因。并使用经典的疾病传播模型SIR模型对这两种算法进行了信息传播的仿真模拟,仿真结果显示LeaderRank算法用于在线社会网络节点重要性排序效果更好。
It is very important that identifying the influential nodes of complex network for information spreading and network marking. PageRank and LeaderRank algorithm are famous in ranking important nodes of complex network. Pokec is the most popular on-line social network in Slovakia. The nodes of the network stand for Pokec＇s users,while edges represent the links between users. Each node＇s PageRank value and LeaderRank value is computed out while the users＇ descending order by that value is obtained.Comparing with degree centrality,there are advantages and disadvantages of these two algorithms. Taking top-N influential users as seeds, information spreading is simulated on SIR model with the two algorithms. The results show that LeaderRank algorithm performs better.
出处
《信息技术》
2015年第4期8-11,共4页
Information Technology
基金
国家自然科学基金(70971089)
上海市一流学科(系统科学)项目资助(XTKX2012)
上海市研究生创新基金项目(JWCXSL1302)
