Title: Dual $φ$-divergences estimation in normal models

Abstract: A class of robust estimators which are obtained from dual representation of $\phi$-divergences, are studied empirically for the normal location model. Members of this class of estimators are compared, and it is found that they are efficient at the true model and offer an attractive alternative to the maximum likelihood, in term of robustness .