兰州大学机构库 >数学与统计学院
基于随机森林的变量选择方法研究
Alternative TitleStudies on the Variable Selection Methods Based on Random Forests
卢天驰
Subtype学士
Thesis Advisor李周平
2021-05-20
Degree Grantor兰州大学
Place of Conferral兰州
Degree Name理学学士
Degree Discipline数学与应用数学
Keyword变量选择 随机森林 正则化惩罚方法
Abstract在现代统计模型中, 变量选择问题因为可以增强模型解释率, 减小计算量而一直备受关注. 本文基于随机森林对变量选择进行研究, 同时将随机森林与正则化方法进行比较, 旨在通过数值模拟和数据分析的方法来综合比较这两类变量选择方法, 具体研究步骤如下:第一, 对随机森林及正则化惩罚方法相关理论进行了阐述. 第二, 在线性和Logistic 分类模型中进行模拟, 比较出各个方法的适用性和有效性. 第三, 应用这些方法对一些实际数据进行了分析. 最终, 本文通过模拟和数据分析得到了各种方法在不同数据集上的优劣, 特点.
Other AbstractIn modern statistical models, the problem of variable selection has been paid much attentionbecause it can enhance the explanatory rate of models and reduce the amount of calculation. Thisthesis studies variable selection based on random forests, and compares random forests with regularization penalty methods. The specifific research steps are as follows: First, the related theoriesof random forests and regularization penalty methods are explained.The applicability and validityof each method are compared by simulation in linear and Logistic classifification models. Third,some actual data are analyzed by these methods. Finally, the advantages and disadvantages ofeach method on different datasets are obtained through simulation and data analysis.
Pages42
URL查看原文
Language中文
Document Type学位论文
Identifierhttps://ir.lzu.edu.cn/handle/262010/460975
Collection数学与统计学院
Affiliation
数学与统计学院
First Author AffilicationSchool of Mathematics and Statistics
Recommended Citation
GB/T 7714
卢天驰. 基于随机森林的变量选择方法研究[D]. 兰州. 兰州大学,2021.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Altmetrics Score
Google Scholar
Similar articles in Google Scholar
[卢天驰]'s Articles
Baidu academic
Similar articles in Baidu academic
[卢天驰]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[卢天驰]'s Articles
Terms of Use
No data!
Social Bookmark/Share
No comment.
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.