Comparison of Logistic Regression and Machine Learning Approaches in Predicting Depressive Symptoms: A National-Based Study.

IF 1.8 4区医学 Q3 PSYCHIATRY

Psychiatry Investigation Pub Date : 2025-03-01 Epub Date: 2025-03-18 DOI:10.30773/pi.2024.0156

Xing-Xuan Dong, Jian-Hua Liu, Tian-Yang Zhang, Chen-Wei Pan, Chun-Hua Zhao, Yi-Bo Wu, Dan-Dan Chen

{"title":"Comparison of Logistic Regression and Machine Learning Approaches in Predicting Depressive Symptoms: A National-Based Study.","authors":"Xing-Xuan Dong, Jian-Hua Liu, Tian-Yang Zhang, Chen-Wei Pan, Chun-Hua Zhao, Yi-Bo Wu, Dan-Dan Chen","doi":"10.30773/pi.2024.0156","DOIUrl":null,"url":null,"abstract":"Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic.Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC).Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models.Conclusion: Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.","PeriodicalId":21164,"journal":{"name":"Psychiatry Investigation","volume":"22 3","pages":"267-278"},"PeriodicalIF":1.8000,"publicationDate":"2025-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11962532/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Psychiatry Investigation","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.30773/pi.2024.0156","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/3/18 0:00:00","PubModel":"Epub","JCR":"Q3","JCRName":"PSYCHIATRY","Score":null,"Total":0}

引用次数: 0

Abstract

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic.

Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC).

Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models.

Conclusion: Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

查看原文本刊更多论文

预测抑郁症状的逻辑回归和机器学习方法的比较：一项基于国家的研究。

目的：据报道，机器学习（ML）比传统的统计技术具有更好的预测能力。本研究的目的是评估ML算法和逻辑回归（LR）在预测COVID-19大流行期间抑郁症状的有效性。方法：在一项涉及21,916名参与者的全国性横断面研究中进行分析。本研究的机器学习算法包括随机森林（RF）、支持向量机（SVM）、神经网络（NN）和梯度增强机（GBM）方法。性能指标为灵敏度、特异度、准确度、精密度、f1评分、受试者工作特征曲线下面积（AUC）。结果：LR和NN在auc方面表现最好。除RF外，大多数ML模型的过拟合风险可以忽略不计，GBM获得最高的敏感性、特异性、准确性、精密度和f1评分。因此，LR、NN和GBM模型名列最佳模型之列。结论：与ML模型相比，LR模型在预测抑郁症状和识别潜在危险因素方面的效果与ML模型相当，但其过拟合风险也较低。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Psychiatry Investigation PSYCHIATRY-

CiteScore

4.10

自引率

3.70%

发文量

105

审稿时长

6-12 weeks

期刊介绍： The Psychiatry Investigation is published on the 25th day of every month in English by the Korean Neuropsychiatric Association (KNPA). The Journal covers the whole range of psychiatry and neuroscience. Both basic and clinical contributions are encouraged from all disciplines and research areas relevant to the pathophysiology and management of neuropsychiatric disorders and symptoms, as well as researches related to cross cultural psychiatry and ethnic issues in psychiatry. The Journal publishes editorials, review articles, original articles, brief reports, viewpoints and correspondences. All research articles are peer reviewed. Contributions are accepted for publication on the condition that their substance has not been published or submitted for publication elsewhere. Authors submitting papers to the Journal (serially or otherwise) with a common theme or using data derived from the same sample (or a subset thereof) must send details of all relevant previous publications and simultaneous submissions. The Journal is not responsible for statements made by contributors. Material in the Journal does not necessarily reflect the views of the Editor or of the KNPA. Manuscripts accepted for publication are copy-edited to improve readability and to ensure conformity with house style.