Improved mutual information measure for clustering, classification, and community detection
Improved mutual information measure for clustering, classification, and community detection
复制标题
DOI：
10.1103/physreve.101.042304
复制
发表时间：
2020-04-23
影响因子：
2.4
通讯作者：
Young, Jean-Gabriel
中科院分区：
文献类型：
--
作者：
Newman, M. E. J.;Cantwell, George T.;Young, Jean-Gabriel
被称为互信息的信息理论度量被广泛用作量化同一对象集合的两个不同标记或划分的相似性的方式，例如在机器学习中的聚类和分类问题或网络科学中的社区检测问题中出现。在这里，我们认为，标准的互信息，通常定义，省略了一个关键的术语，可以在现实世界的条件下变得很大，产生的结果，可以大大错误。我们推导出这个缺失项的表达式，从而写出一个校正的互信息，即使在标准测量失败的情况下也能给出准确的结果。我们将讨论新措施的实际实施，并给出示例应用程序。
The information theoretic measure known as mutual information is widely used as a way to quantify the similarity of two different labelings or divisions of the same set of objects, such as arises, for instance, in clustering and classification problems in machine learning or community detection problems in network science. Here we argue that the standard mutual information, as commonly defined, omits a crucial term which can become large under real-world conditions, producing results that can be substantially in error. We derive an expression for this missing term and hence write a corrected mutual information that gives accurate results even in cases where the standard measure fails. We discuss practical implementation of the new measure and give example applications.