华中科技大学学报（自然科学版）Issue(5)：104-108,5.DOI:10.13245/j.hust.150520
一种互联网拓扑结构局部特征度量修正算法
Correcting algorithm for local characteristic metrics of Internet topology structure
摘要
Abstract
To solve the problem of analysis deviation for the characteristic metrics caused by the heter‐ogeneity and low accuracy in the IP geolocation databases ,a correcting algorithm (IP magnifying glass ,IPMG) based on machine learning was proposed in the research of Internet topology structure . Two new local characteristic metrics ,location degree and location betweenness ,were defined ,with the combination of huge traceroute data and the IP geographical location information in the IP geoloca‐tion databases .The power law of the two characteristic metrics and the relationships between them and IP geographic location were analyzed .Machine learning method was used to correct the analysis deviation for the characteristic metrics in different geolocation databases .The effective of the IPMG method was evaluated by both of the cross validation method and landmark validation method .The re‐sults show that the IPMG method makes an effective correction for the characteristic metrics and in‐creases the accuracy of the IP geolocation databases .关键词
复杂网络/拓扑结构/特征度量/幂律分布/机器学习Key words
complex networks/topology structure/characteristic metric/power law/machine learn-ing分类
信息技术与安全科学引用本文复制引用
王婷,许可,王娜,宋俊德..一种互联网拓扑结构局部特征度量修正算法[J].华中科技大学学报（自然科学版）,2015,(5):104-108,5.基金项目
国家科技支撑计划资助项目（2014BA H26F02）． （）
