A class of generalized Shannon-McMillan theorems for arbitrary discrete information source (Q2889936)

This paper studies a class of strong limit theorems (generalized Shannon-McMillan theorems) for the relative entropy densities of random sums of an arbitrary information source, with a countable state space, and for the generalized entropy density of random sums. These results extend some similar results in [\textit{W. Liu} and \textit{W. Yang}, Stochastic Processes Appl. 61, No. 1, 129--145 (1996; Zbl 0861.60042); \textit{W. Yang, Z. Ye} and \textit{W. Liu}, ``A local convergence theorem for partial sums of stochastic adapted sequences'', Czech. Math. J. 56, No. 2, 525--532 (2006; Zbl 1164.60338].