{"title":"机器学习分类算法的比较及在弗雷明汉心脏研究中的应用","authors":"Nabil Kahouadji","doi":"arxiv-2402.15005","DOIUrl":null,"url":null,"abstract":"The use of machine learning algorithms in healthcare can amplify social\ninjustices and health inequities. While the exacerbation of biases can occur\nand compound during the problem selection, data collection, and outcome\ndefinition, this research pertains to some generalizability impediments that\noccur during the development and the post-deployment of machine learning\nclassification algorithms. Using the Framingham coronary heart disease data as\na case study, we show how to effectively select a probability cutoff to convert\na regression model for a dichotomous variable into a classifier. We then\ncompare the sampling distribution of the predictive performance of eight\nmachine learning classification algorithms under four training/testing\nscenarios to test their generalizability and their potential to perpetuate\nbiases. We show that both the Extreme Gradient Boosting, and Support Vector\nMachine are flawed when trained on an unbalanced dataset. We introduced and\nshow that the double discriminant scoring of type I is the most generalizable\nas it consistently outperforms the other classification algorithms regardless\nof the training/testing scenario. Finally, we introduce a methodology to\nextract an optimal variable hierarchy for a classification algorithm, and\nillustrate it on the overall, male and female Framingham coronary heart disease\ndata.","PeriodicalId":501340,"journal":{"name":"arXiv - STAT - Machine Learning","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Comparison of Machine Learning Classification Algorithms and Application to the Framingham Heart Study\",\"authors\":\"Nabil Kahouadji\",\"doi\":\"arxiv-2402.15005\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The use of machine learning algorithms in healthcare can amplify social\\ninjustices and health inequities. While the exacerbation of biases can occur\\nand compound during the problem selection, data collection, and outcome\\ndefinition, this research pertains to some generalizability impediments that\\noccur during the development and the post-deployment of machine learning\\nclassification algorithms. Using the Framingham coronary heart disease data as\\na case study, we show how to effectively select a probability cutoff to convert\\na regression model for a dichotomous variable into a classifier. We then\\ncompare the sampling distribution of the predictive performance of eight\\nmachine learning classification algorithms under four training/testing\\nscenarios to test their generalizability and their potential to perpetuate\\nbiases. We show that both the Extreme Gradient Boosting, and Support Vector\\nMachine are flawed when trained on an unbalanced dataset. We introduced and\\nshow that the double discriminant scoring of type I is the most generalizable\\nas it consistently outperforms the other classification algorithms regardless\\nof the training/testing scenario. Finally, we introduce a methodology to\\nextract an optimal variable hierarchy for a classification algorithm, and\\nillustrate it on the overall, male and female Framingham coronary heart disease\\ndata.\",\"PeriodicalId\":501340,\"journal\":{\"name\":\"arXiv - STAT - Machine Learning\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-02-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - STAT - Machine Learning\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2402.15005\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - STAT - Machine Learning","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2402.15005","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
在医疗保健领域使用机器学习算法可能会扩大社会不公正和健康不平等。在问题选择、数据收集和结果定义过程中,偏差的加剧可能会发生和加剧,而本研究涉及的是机器学习分类算法开发和部署过程中出现的一些可推广性障碍。以弗雷明汉冠心病数据为例,我们展示了如何有效选择概率截断点,将二分变量的回归模型转换为分类器。然后,我们比较了八种机器学习分类算法在四种训练/测试情景下预测性能的抽样分布,以检验它们的普适性及其延续偏差的可能性。我们发现,当在不平衡数据集上进行训练时,极梯度提升算法和支持向量机都存在缺陷。我们介绍并展示了 I 型双判别评分法的通用性最强,因为无论在何种训练/测试场景下,它的表现都始终优于其他分类算法。最后,我们介绍了一种为分类算法提取最佳变量层次结构的方法,并在总体、男性和女性弗雷明汉冠心病数据中进行了演示。
Comparison of Machine Learning Classification Algorithms and Application to the Framingham Heart Study
The use of machine learning algorithms in healthcare can amplify social
injustices and health inequities. While the exacerbation of biases can occur
and compound during the problem selection, data collection, and outcome
definition, this research pertains to some generalizability impediments that
occur during the development and the post-deployment of machine learning
classification algorithms. Using the Framingham coronary heart disease data as
a case study, we show how to effectively select a probability cutoff to convert
a regression model for a dichotomous variable into a classifier. We then
compare the sampling distribution of the predictive performance of eight
machine learning classification algorithms under four training/testing
scenarios to test their generalizability and their potential to perpetuate
biases. We show that both the Extreme Gradient Boosting, and Support Vector
Machine are flawed when trained on an unbalanced dataset. We introduced and
show that the double discriminant scoring of type I is the most generalizable
as it consistently outperforms the other classification algorithms regardless
of the training/testing scenario. Finally, we introduce a methodology to
extract an optimal variable hierarchy for a classification algorithm, and
illustrate it on the overall, male and female Framingham coronary heart disease
data.