Regularization, penalization techniques and local modelling
2018-07-02
Cycle of 5 seminars held by Professor Irene Gijbels (Department of Mathematics, Section of Statistics at the University of Leuven) at the IAC seminar hall in Rome, according to the following schedule: - Monday 2/07/2018 10:00 - 13:00 - Tuesday 3/07/2018 9:00 - 12:00 - Wednesday 4/07/2018 9:00 - 11:00 - Thursday 5/07/2018 9:00 - 11:00
(I). Multiple linear regression: regularization and penalization methods
* least squares regression and Ridge regression
* regularization and penalization techniques
(II). Flexible mean regression: regularization and penalizations methods
* penalized likelihood regression for generalized linear models
* additive regression models and penalization techniques
* varying coefficient models and penalization techniques
* grouped regularization methods.
(III). Regularization and penalizations methods in mean and dispersion regression
* mean and dispersion estimation and P-splines approximations
* mean and dispersion estimation in a framework of proper dispersion models.
(IV). Nonparametric mean regression using local polynomial fitting
* from parametric polynomial regression to local polynomial regression
* asymptotic properties of local polynomial regression
* selection of procedure parameters
(V). Local likelihood methods and nonparametric logistic regression
(VI). Elements of nonparametric inference for functional data analysis.
