期刊
PSYCHIATRY RESEARCH
卷 306, 期 -, 页码 -出版社
ELSEVIER IRELAND LTD
DOI: 10.1016/j.psychres.2021.114261
关键词
Depression; Machine learning; Middle-aged and elderly; NHANES
类别
资金
- National Key Research and Development Program of China [2017YFC0108602]
The study used five machine learning models to evaluate depression among middle-aged and elderly participants, with CatBoost performing the best in identifying depression. LR and SVM ranked second, while BP and RF models had slightly worse performance.
Depression is one of the most common mental health problems in middle-aged and elderly people. The establishment of risk factor-based depression risk assessment model is conducive to early detection and early treatment of high-risk groups of depression. Five machine learning models (logistic regression (LR); back propagation (BP); random forest (RF); support vector machines (SVM); category boosting (CatBoost) were used to evaluate the depression among 8374 middleaged people and 4636 elderly people in the NHANES database from 2011 to 2018. In the 2011-2018 cycle, the estimated prevalence of depression was 8.97% in the middle-aged participants and 8.02% in the elderly participants. Among the middle-aged and elderly participants, CatBoost was the best model to identify depression, and its area under the working characteristic curve (AUC) reaches the highest. The second is LR model and SVM model, while the performance of BP and RF model was slightly worse. The primary influencing factor of depression in middle-aged male is alanine aminotransferase. All five machine learning models can identify the occurrence of depression in the NHANES data set through social demographics, lifestyle, laboratory data and other data of middle-aged and elderly people, and among five models, the CatBoost model performed best.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据