Robust predictive distribution and bias-calibration in Linear Models

Diplôme Thèse Robust predictive distribution and bias-calibration in Linear Models eng 2019-12-20 author Branca Mattia 724309 director Cantoni Eva 0000-0002-6614-7772 93895 director Ronchetti Elvezio 0000-0003-2231-4556 22756 10.13097/archive-ouverte/unige:131077 urn:nbn:ch:unige-1310776 GSEM 74 This thesis focuses on the concept of predictive distributions and bias calibration. At first, an extension of the concept of predictive distributions under contamination is studied in the case of Generalized Linear Models. A sensitivity analysis of the impact of contamination on the predictive distribution is studied making use of the class of $M$-estimators. In a second step, based on the available literature on bias-calibrated estimation in linear regression, the bases to implement bias-calibration for predictive distributions are studied and developed. This development is based on the finite-sample setting and an important aspect of the reasoning behind this contribution is the distinction between representative and non-representative outliers. As a result of this distinction, the use of the bias-calibration approach allows to integrate information from representative outliers within the predictive distribution. Predictive distribution Bias-calibration Robust statistics Laplace approximation Dewey 310 117 Research Center for Statistics S5 false b27f76a3-6389-4326-9d06-41ea2e1dfb0b Thesis Public gsemthesis_manuscriptMB_Dec19.pdf 1137193 application/pdf openAccess Faculté d'économie et de management / Research Center for Statistics

Title	Published in	Access level	OA Policy	Year	Views	Downloads

Robust predictive distribution and bias-calibration in Linear Models				2019	629	401

Archive ouverte UNIGE

Branca, Mattia