Abstract:
We consider a power series at a fixed point $\rho \in (0.5,1)$, where random
coefficients assume a value $0$ or $1$ and form a stationary ergodic aperiodic
process. The Erdős measure is the distribution law of such a series. The
problem of absolute continuity of the Erdős measure is reduced to the
problem of determining when the corresponding hidden Markov chain is a Parry–Markov
chain. For the golden ratio and a 1-Markov chains, we give necessary and
sufficient conditions for absolute continuity of the Erdős measure and,
using Blackwell–Markov chains, provide a new proof that the necessary
conditions obtained earlier by Bezhaeva and Oseledets [Theory Probab. Appl., 51 (2007), pp. 28–41] are also sufficient. For tribonacci numbers and 1-Markov chains, we give a new proof of the theorem on singularity of the Erdős measure. For tribonacci numbers and 2-Markov chains, we find only two cases with absolute continuity.