The diehard test is a statistical test battery to measure the quality of a random number generator. They were developed by George Marsaglia for several years and were first published in 1995 on random number CD-ROMs.
It should be noted that most tests on DIEHARD return a p value, which should be uniform in [0,1] if the input file contains completely independent random bits. The p values ââare obtained by p = F ( X ), where F is the assumption distribution of the random sample variable X - often normal. But it is assumed that F is only an asymptotic assumption, which would be the worst in the tail. Thus you should not be surprised by the value of p occasionally close to 0 or 1, such as 0.0012 or 0.9983. When the bit stream is completely FAILED, you will get p s from 0 or 1 to six or more places. Since there are many tests, it is not possible aa p & lt; 0.025 or p & gt; 0.975 means that the RNG has "failed the test at the 0.05 level". We expect such events to occur among the hundreds of events that DIEHARD produces, even conditioned on random number-makers to be perfect.
Video Diehard tests
See also
- Test randomness
- TestU01
Maps Diehard tests
Note
External links
- "Marsaglia Random Number CDROM includes Diehard Battery of Tests Randomness". Florida State University. 1995. Archived from the original in 2016-01-25.
- Robert G. Brown. "Dieharder: A Random Number Test Suite".
- RÃÆ' à © nyi, AlfrÃÆ'à © d (1953). "On order statistics theory". Acta Mathematica Academiae Scientiarum Hungaricae . Acta Mathematica Hungarica. 4 (3-4): 191. doi: 10.1007/BF02127580.
Source of the article : Wikipedia