Bagging for the highly adaptive lasso Público

Yu, Haoyong (Spring 2020)

Permanent URL: https://etd.library.emory.edu/concern/etds/g445cf348?locale=pt-BR
Published

Abstract

Prediction is a common goal in the statistic world. Many new estimators are created every year to seek better quality of prediction in various situation. A new estimator called highly adaptive lasso estimator was proved to be competitive with other popular machine learning methods and had theoretical advantages. Furthermore, the prediction performance of this estimator may be furthered by combining with some unique methods. Bagging is a common ensemble method that can be utilized to improve the performance of prediction. Feature bagging is a promising usage of traditional bagging method. We propose a new estimator that we call bagged highly adaptive lasso estimator based on feature bagging approach. We show via simulation and public data analysis that our estimator seems not provide more benefits by additional aggregating bootstrap procedures.

Table of Contents

1.      Introduction 1

2.      Methods 2

2.1   Highly adaptive lasso 2

2.2   Bagging 3

2.3   Bagged highly adaptive LASSO 4

3.      Simulation 4

4.      Data Analysis 9

5.      Conclusion 10

References 12

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Subfield / Discipline
Degree
Submission
Language
  • English
Research Field
Palavra-chave
Committee Chair / Thesis Advisor
Última modificação

Primary PDF

Supplemental Files