SearcharxivSearch

arXiv subjects

Jairo. K. Mengue

Publications and source records attributed to Jairo. K. Mengue.

2 recordsLinked to original sources

Large deviation for Gibbs probabilities at zero temperature and invariant idempotent probabilities for iterated function systems

We consider two compact metric spaces $J$ and $X$ and a uniform contractible iterated function system $\{ϕ_j: X \to X \, | \, j \in J \}$. For a Lipschitz continuous function $A$ on $J \times X$ and for each $β>0$ we consider the Gibbs probability $ρ_{_{βA}}$. Our goal is to study a large deviation principle for such family of probabilities as $β\to +\infty$ and its connections with idempotent probabilities. In the non-place dependent case ($A(j,x)=A_j,\,\forall x\in X$) we will prove that $(ρ_{_{βA}})$ satisfy a LDP and $-I$ (where $I$ is the deviation function) is the density of the unique invariant idempotent probability for a mpIFS associated to $A$. In the place dependent case, we prove that, if $(ρ_{_{βA}})$ satisfy a LDP, then $-I$ is the density of an invariant idempotent probability. Such idempotent probabilities were recently characterized through the Mañé potential and Aubry set, therefore we will obtain an identical characterization for $-I$.

math.DS

The generalized IFS Bayesian method and an associated variational principle covering the classical and dynamical cases

We introduce a general IFS Bayesian method for getting posterior probabilities from prior probabilities, and also a generalized Bayes' rule, which will contemplate a dynamical, as well as a non-dynamical setting. Given a loss function ${l}$, we detail the prior and posterior items, their consequences and exhibit several examples. Taking $Θ$ as the set of parameters and $Y$ as the set of data (which usually provides {random samples}), a general IFS is a measurable map $τ:Θ\times Y \to Y$, which can be interpreted as a family of maps $τ_θ:Y\to Y,\,θ\inΘ$. The main inspiration for the results we will get here comes from a paper by Zellner (with no dynamics), where Bayes' rule is related to a principle of minimization of {information.} We will show that our IFS Bayesian method which produces posterior probabilities (which are associated to holonomic probabilities) is related to the optimal solution of a variational principle, somehow corresponding to the pressure in Thermodynamic Formalism, and also to the principle of minimization of information in Information Theory. Among other results, we present the prior dynamical elements and we derive the corresponding posterior elements via the Ruelle operator of Thermodynamic Formalism; getting in this way a form of dynamical Bayes' rule.

math.DS