Model selection and estimation in high dimensional settings

Variable selection;High dimensional statistics;Regularization

Ngueyep Tzoumpe, Rodrigue ; Serban, Nicoleta Industrial and Systems Engineering Xie, Yao Vandekerkhove, Pierre Goldsman, David M. Vengazhiyil, Roshan ; Serban, Nicoleta

University:Georgia Institute of Technology

Department:Industrial and Systems Engineering

关键词: Variable selection;    High dimensional statistics;    Regularization;   

Others  :  https://smartech.gatech.edu/bitstream/1853/53551/1/NGUEYEPTZOUMPE-DISSERTATION-2015.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

Several statistical problems can be described as estimation problem, where the goal is to learn a set of parameters, from some data, by maximizing a criterion. These type of problems are typically encountered in a supervised learning setting, where we want to relate an output (or many outputs) to multiple inputs. The relationship between these outputs and these inputs can be complex, and this complexity can be attributed to the high dimensionality of the space containing the inputs and the outputs; the existence of a structural prior knowledge within the inputs or the outputs that if ignored may lead to inefficient estimates of the parameters; and the presence of a non-trivial noise structure in the data. In this thesis we propose new statistical methods to achieve model selection and estimation when there are more predictors than observations. We also design a new set of algorithms to efficiently solve the proposed statistical models. We apply the implemented methods to genetic data sets of cancer patients and to some economics data.