电信科学 ›› 2016, Vol. 32 ›› Issue (6): 153-162.doi: 10.11959/j.issn.1000-0801.2016169

• 综述 • 上一篇    下一篇

大数据下图三角计算的研究进展

金宏桥,董一鸿   

  1. 宁波大学信息科学与工程学院,浙江 宁波 315211
  • 出版日期:2016-06-20 发布日期:2016-07-20

Research progress of triangle counting in big data

Hongqiao JIN,Yihong DONG   

  1. Faculty of Electrical Engineering and Computer Science,Ningbo University,Ningbo 315211,China
  • Online:2016-06-20 Published:2016-07-20

摘要:

图三角数量的计算是计算网络聚集系数和传递性的重要步骤,广泛应用于重要角色识别、垃圾邮件检测、社区发现、生物检测等。在大数据背景下,计算图中三角形算法主要面临时空消耗和计算准确性两大难题。介绍了代表性的大图中计算三角形的算法,主要存在准确计算和近似计算两大类。准确计算算法又分为内存算法、外存算法和分布式算法,时空消耗或I/O消耗很大。近似计算算法中,有辅助算法、非流式算法和流式算法之分。最后对计算三角形算法进行了归纳总结。

关键词: 准确计算, 近似计算, 三角形,

Abstract:

Counting triangles in a graph is an important step to calculate the clustering coefficient and the transitivity ratio of the network,which is widely used in important role identification,spam detection,community discovery,biological detection etc.Counting triangles algorithm is mainly faced with two major problems of space-time consumption and accuracy.The representative algorithm of the counting triangles in the big graph was introduced.There existed two kinds of algorithms,which were exact counting algorithm and approximate counting algorithm.Exact counting algorithms were divided into internal memory algorithm,external memory algorithm and distributed algorithm.The space-time consumption or I/O consumption of exact counting algorithm was very large.Approximate counting algorithms were divided into auxiliary algorithm,static algorithm and streaming algorithm.In the end,the counting triangles algorithms were summarized.

Key words: exact counting, approximate counting, triangle, graph

No Suggested Reading articles found!