| Group Lasso 的应用研究分析 |
Alternative Title | Research on Group Lasso and Its Application
|
| 陈玉雯 |
Thesis Advisor | 赵学靖
|
| 2013-05-23
|
Degree Grantor | 兰州大学
|
Place of Conferral | 兰州
|
Degree Name | 学士
|
Keyword | 模型选择
Lasso
Group Lasso
LARS
|
Abstract | 统计回归分析中,模型选择和参数估计一直是一个热点问题。近几年来,对Lasso方法的研究非常广泛,Group Lasso 方法是Lasso方法的一种衍生。本文以线性模型为切入点,将Lasso 方法、Group Lasso方法应用于加速寿命模型,并对模型系数进行压缩估计。利用最小角回归算法(LARS)对估计方法进行求解,并且编译了matlab 程序实现变量选择。实验结果表明, 当变量存在明显分组属性时,Group Lasso 的估计方法可以更好地应用于分组变量选择的模型中,可以更好地达到模型的解释性和参数估计的准确性。 |
Other Abstract | Model selection and parameter estimation attached intensively interesting in statistical regression analysis. In recent years, the Lasso has a very wide range of applications, Group lasso is a natural extension of lasso and selects variables in a grouped manner. In this paper, the shrinkage methods, Lasso, Group Lasso was investigated to the multi-factor linear model and shrinking the coefficient, which can be solved by Least Angle regression algorithm. Furthermore Matlab code are explored to compared Lasso and Group Lasso. In high-dimensional with group covariates, Group Lasso has better performance, which can achieve better explanatory and the accuracy of parameter. |
URL | 查看原文
|
Language | 中文
|
Document Type | 学位论文
|
Identifier | https://ir.lzu.edu.cn/handle/262010/225706
|
Collection | 数学与统计学院
|
Recommended Citation GB/T 7714 |
陈玉雯. Group Lasso 的应用研究分析[D]. 兰州. 兰州大学,2013.
|
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.