2015/06/06 by Diaa Al Mohamad, Mohamad, Diaa Al
Mathematics · Physics and Astronomy · #62F35 #Advanced Statistical Methods and Models #FOS: Computer and information sciences #Methodology (stat.ME) #Statistical Distribution Estimation and Applications #Statistical Mechanics and Entropy #msc:62F35 #stat.ME
paper · pdf · doi:10.48550/arxiv.1506.02166
Working paper. It only lacks some revisions before submitting
openalex publication_date 2015/06/06 · arxiv created 2015/10/10 · arxiv updated 2015/10/13 · openalex created_date 2022/10/04 · openalex updated_date 2026/07/28
The aim of this paper is to study different estimation procedures based on φ-divergences. The dual representation of φ-divergences based on the Fenchel-Legendre duality is the main interest of this study. It provides a way to estimate φ-divergences by a simple plug-in of the empirical distribution without any smoothing technique. Resulting estimators are thoroughly studied theoretically and with simulations showing that the so called minimum φ-divergence estimator (MDφDE) is generally non robust and behaves similarly to the maximum likelihood estimator. We give some arguments supporting the non robustness property, and give insights on how to modify the classical approach. An alternative class of φ-divergences robust estimators based on the dual representation is presented. We study consistency and robustness properties from an influence function point of view of the new estimators. In a second part, we invoke the Basu-Lindsay approach for approximating φ-divergences and provide a comparison between these approaches. The so called dual φ-divergence is also discussed and compared to our new estimator. A full simulation study of all these approaches is given in order to compare efficiency and robustness of all mentioned estimators against the so-called minimum density power divergence, showing encouraging results in favor of our new class of minimum dual φ-divergences.