High-Efficiency Min-Entropy Estimation Based on Neural Network for Random Number Generators

Table 7

Entropy estimates for time-varying data.

Data class	Correct	90B’s predictors				Our work
		MCW	Lag	MultiMMC	LZ78Y	FNN	RNN
gradual(0.2)	0.6345	0.5290	0.7808	0.7240	0.7790	0.6288	0.6289
gradual(0.3)	0.7437	0.7378	0.9221	0.8416	0.9243	0.7430	0.7460
gradual(0.4)	0.8645	0.8631	0.9786	0.9518	0.9739	0.8648	0.8637
period(0.2)	0.6345	0.5537	0.7428	0.5537	0.7669	0.6205	0.6209
period(0.3)	0.7437	0.7393	0.9218	0.8476	0.9233	0.7377	0.7375
period(0.4)	0.8645	0.8639	0.9767	0.9632	0.9796	0.8653	0.8632
sudden(0.2)	0.3219	0.3203	0.4663	0.3386	0.4484	0.3217	0.3229
sudden(0.3)	0.5146	0.5110	0.5857	0.9984	0.7663	0.5110	0.5119
sudden(0.4)	0.7370	0.7338	0.8699	0.9984	0.9389	0.7339	0.7345