NEW STATISTICAL BINARY RANDOM TEST

Abstract

Binary streams1 can be tested, using standard tests2. The efficiency of some tests is not remarkable big, because they do not examine some stream features, which are important to be verified [for example, mono-bit test satisfies every stream that consists of approximate one half zeros (i.e. ones), no matter if all of them are consecutive or alternate), and so several tests are to be combined. In our proposed test, distribution of frequencies of blocks3 and gaps4 are determined and for given number of the blocks and the gaps expected distribution of frequencies of the blocks and the gaps is expected. 2 is calculated, and it is checked whether the probability of calculated 2 is above the threshold  of significance. In the same way both distributions (blocks and gaps) are evaluated.

Keywords

STATISTICAL