新浪微博snow ball采样算法偏差研究(英文)  

Bias of snow ball sampling on Sina-Microblog

作　　者：李晓丽[1] 许可[1] 宋俊德[1] 

机构地区：[1]北京邮电大学,北京100876

出　　处：《软件》2012年第11期228-230,共3页Software

摘　　要：随着在线社交网络的广泛应用,多种多样的采样算法应用于在线社交网络结构和特性的建模和分析。本文首先对这些常用的在线社交网络采样算法做分类对比,然后对应用相对广泛的snow ball采样算法进行深入研究。本文通过对国内最早、应用最广泛的在线社交网络新浪微博的数据进行采样分析,研究新浪微博网络结构的特点,验证得出数据量不足会导致在线社交网络的网络半径和度分布产生较大偏差。With the increase population of Online Social Networks, various graph-sampling algorithms are proposed to model, analysis and implement the structures and characteristics of OSNs. In this thesis, we make a contrast on the bias of the sampling algorithms. In order to take an insight into the bias of snow ball sampling, we construct an experiment on Sina-Microblog, which is the earliest and most popular microblogging service. Our data suggests that small dataset usually causes a bigger bias on degree distribution and diameter of the OSNs network.

关 键 词：在线社交网络 采样算法 偏差 度分布 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]