Optimal estimation of parameters in two-component contamination mixtures

Jeudi 26 janvier 2017 14:00-15:00 - Clément Marteau - Lyon 1

Résumé : We consider a parametric density contamination model. We work with a sample of i.i.d. data having a common density $f^\star =(1-\lambda^\star) \phi + \lambda^\star \phi(.-\mu^\star)$ where the shape $\phi$ is assumed to be known. We establish optimal rates of convergence for the estimation of the mixture parameters $(\lambda^\star,\mu^\star)$. In particular, we prove that the classical parametric rate $1/n$ can not be reached when at least one of these parameters is allowed to tend to $0$ with $n$.

Lieu : Salle 117-119

Optimal estimation of parameters in two-component contamination mixtures  Version PDF