Empirical assessment of efficiency of entropy source for random number generators using autocorrelation factor test

To achieve true randomness of numbers, the entropy source needs to be efficient. The statistical testing of efficiency of entropy source is necessary as it is used in random number generators for cryptography, simulation, statistical sampling, etc. 5000 numbers each were generated in four experimental entropy conditions: Control experiment with white light, Experimental conditions with multicoloured light with light path not disrupted, while light with path disrupted by thermocol balls, multicoloured light with path disrupted by thermocol balls. The numbers thus generated using each entropy condition were tested by Auto Correlation Factor Test to identify optimal experimental condition. The results were displayed in graphical and tabular format. The four graphs depicted the autocorrelations clearly showing that the best results were obtained for the fourth experimental condition ‘multicoloured light fan on’. It was found that Autocorrelation factor test is a powerful test for empirical assessment of efficiency of entropy source for random number generation. If the random number generator developed by any individual or organization is tested by Autocorrelation factor test, the efficiency of the entropy source can be pre-established thus preventing post hazards like hacking of random numbers. The security of data is a major concern in all areas such as defense, banking, research, designing and evaluation of examinations, etc.