There exist a lot of randomness test methods, and most of them have parameters. As it is not practical to do all randomness tests in practice, it is important to study the relations among these methods. In this paper, four kinds of relations between randomness tests and a conception of "correlation degree of randomness tests" are defined firstly based on the statistics theory firstly. And the correlation degree is measured by means of the entropy method. Then, the relevancies between the four relations and the correlation degree are proved. And an algorithm of calculating correlation degree and a selection policy are provided as well. The work of this paper is helpful for selecting reasonable and scientific randomness tests and parameters. In addition, the randomness test methods adopted by NIST (National Institute of Standards and Technology) in AES (advanced encryption standard) are explored by using the correlation degree and some dependence relations are found.%目前,存在众多的随机性检测项目,并且许多项目都带有参数.选择所有的项目进行检测不现实,因此需要研究检测项目之间的关系.从统计学角度出发,对检测项目的相关性进行研究,首先定义了检测项目之间存在的4种关系,提出了检测项目相关度的概念,然后利用熵值法对检测项目相关度进行度量,并证明了这4种关系与相关度的联系,同时给出了一种计算相关度的算法和一个基于相关度的检测项目选择策略.所研究的结果为随机性检测项目及其参数选择提供了理论依据.与此同时,利用相关度对NIST在评选AES中所采用的检测项目进行研究,发现了其中一些检测项目之间存在着依赖关系.
展开▼
