医学多元统计方法课程模拟案例教学研究

doi:10.3969/j.issn.1006-5253.2024.01.012

Abstract
Figure/Table
References (0)
Related Citation (7)

Download: PDF (4932 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS) Supporting Info

Guide

Abstract Objective The complexity of the concept of multivariate statistics can easily lead to difficulties in students′ understanding. This article uses simulation experiments to examine the impact of various characteristics of data on the analysis results, in order to strengthen students′ understanding of the principles of multivariate statistics and facilitate their correct application of multivariate statistical methods.Methods Six simulation experiments were designed, and R 4.2.2 was used to generate simulation data and conduct statistical analysis to explore the impact of data characteristics on the 6 statistical analysis methods.Results According to the statistical analysis results of simulation experiments, the multicollinearity among independent variables in multilinear regression will make the regression coefficient estimation of the regression equation inaccurate. The correlation structure between variables in principal component and factor analysis affects the dimensionality reduction effect. In logistic regression, the nonlinear influence of independent variables on dependent variables should be considered, otherwise the results will affect the accuracy of the estimation results. When the proportion of deletion in Cox regression is too large, especially after reaching 60%, the results of the Cox regression model have a significant deviation. The correlation structure between variables affects the results of canonical correlation analysis. When there is a large difference in the proportion of categories, Bayesian discrimination and logistic regression are more appropriate for discrimination, because Bayesian discrimination with prior probability is more flexible.Conclusion The application of multivariate statistical methods needs to fully consider applicable conditions such as data characteristics. Through simulation experiments, students are encouraged to independently run simulation experiments, intuitively understand abstract multivariate concepts, and facilitate the understanding and application of multivariate statistical methods.

Key words： simulation experiment multivariable statistics conceptual principle teaching

Received: 08 November 2023

PACS:

Fund:

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Wang Yaru
	Wang Jiu
	Liu Haixia
	Hou Jing
	Zhang Li
	Li Zhi
	Sun Hongwei

Cite this article:

Wang Yaru,Wang Jiu,Liu Haixia, et al. Research on simulated case teaching in the course of Medical Multiple Statistical Methods [J]. journal1, 2024, 31(1): 61-66.

URL:

http://manu58.magtech.com.cn/Jwk_zgyytj/EN/10.3969/j.issn.1006-5253.2024.01.012 OR http://manu58.magtech.com.cn/Jwk_zgyytj/EN/Y2024/V31/I1/61

［1］杨万龄，李婷婷，罗盈怡，等．“互联网+” 背景下医学统计学混合教学模式的构建［J］．卫生职业教育，2020，38(10)：94-95．

［2］李丽霞，郜艳晖，周舒冬，等．医学多元统计分析课程的教学改革与实践研究［J］．基础医学教育，2018，20(9)：747-750．DOI: 10.13754/j.issn2095-1450.2018.09.12.
［3］达娃玉珍．模拟实验在统计与概率教学中的应用［J］．集成电路应用，2021，38(8)：284-285．DOI: 10.19339/j.issn.1674-2583.2021.08.133.

［4］肖霞，伍兴国．线性回归中多重共线性的几何解释［J］．统计与决策，2021，37(21)：46-51．DOI: 10.13546/j.cnki.tjyjc.2021.21.009.

［5］李红梅，吴喜之，王涛．基于纵向数据与多重共线性数据的神经网络与传统方法比较［J］．统计与决策，2020，36(9)：22-25．DOI: 10.13546/j.cnki.tjyjc.2020.09.004.

［6］孟金玉，袁永生．基于协方差消除多重共线性的一种方法［J］．计算技术与自动化，2021，40(4)：42-46．DOI: 10.16339/j.cnki.jsjsyzdh.202104008.

［7］孙红卫.数据特征对多元统计方法结果影响的模拟试验设计与实现［CP/OL］.https://github.com/hwsun2000/Simulation-experiment-for-multivariate-statisticalmethods-teaching,2023.

［8］何志锋，王斌会．稳健典型相关方法在金融分析中的应用比较研究［J］．数理统计与管理，2021，40(5)：901-913．DOI: 10.13860/j.cnki.sltj.20210305-009.

［9］杨旻，汪吉梅．产妇年龄对新生儿不良结局影响的回顾性队列研究［J］．中国循证儿科杂志，2019，14(4)：276-281．DOI: 10.3969/j.issn.1673-5501.2019.04.006.

［10］NICK T G, CAMPBELL K M. logistic regression［J］. Methods Mol Biol Clifton N J, 2007, 404:273-301. DOI: 10.1007/978-1-59745-530-5_14.

［11］张川．非线性关系的研究及处理［J］. 广西质量监督导报,2020(8):214-215.

［12］蔡丽馨，仲子航，杨旻，等．指定删失比例的生存数据模拟及R实现［J］．中国卫生统计，2022，39(1)：143-148．DOI: 10.3969/j.issn.1002-3674.2022.01.035.

［13］肖媛媛，陈莹，何利平，等．不同删失比例下AFT模型与Cox模型表现比较的模拟研究［J］．中国卫生统计，2017，34(4)：676-680．

［14］董小刚，刘新蕊．二分类响应变量下判别分析与logistic回归比较［J］．长春工业大学学报，2019，40(5)：441-445．DOI: 10.15923/j.cnki.cn22-1382/t.2019.5.04.［15］任涛，吕筠，余灿清，等．疫情后时代我国公共卫生教育和人才队伍建设的思考［J］．中华预防医学杂志，2020，54(5)：457-464．DOI: 10.3760/cma.j.cn112150-20200330-00473.

［16］孙鹏哲，刘宇菲，吕世杰，等．多元统计分析课程的混合教学模式探析［J］．内蒙古农业大学学报(社会科学版)，2019，21(4)：40-44．

［17］张瑞，高铖铖，陈锡程，等．公共卫生硕士专业学位研究生培养质量的调查研究［J］．中国卫生统计，2021，38(1)：10-13. DOI: 10.3969/j.issn.1002-3674.2021.01.003.

［18］周跃进．“多元统计分析” 课程 “翻转课堂+课程思政” 教学模式探索［J］．安徽理工大学学报(社会科学版)，2021，23(5)：96-101．DOI: 10.3969/j.issn.1672-1101.2021.05.017.