Journal of affective disorders · 2025
The risk factors for relapse behavior in individuals with substance use disorders: An interpretable machine learning study.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Xu Jisheng, Li Xue, Liang Xin, Zhang Xuejie, Ying Guosi, Liu Jun
Аннотация
BACKGROUND: Substance abuse has become a serious public health problem worldwide, and finding effective prevention and treatment strategies is undoubtedly an urgent need. This study addresses the risk factors that lead to relapse behaviors among substance use disorders (SUDs) and provides a reference for developing appropriate prevention and treatment strategies for SUDs.
METHODS: In this study, 5255 SUDs were collected by cluster-wide random sampling, including basic personal information, personal health, social relationships, drug abuse behaviors and characteristics, and other information collected to predict the recurrence rate. In segmenting the dataset for training and evaluating several widely used machine learning methods, the best performing eXtreme Gradient Boosting (XGboost) model is explained at different levels by SHapley Additive explanation (SHAP).
RESULTS: 1) The XGboost model achieved the best performance with an area under the receiver operating characteristic curve of 0.76 in the validation set. 2) Ranking of risk factors using SHAP found that drug use behaviors and characteristics were important features for the prediction of relapse behavior. 3) SHAP predicted and interpreted individual recurrence probabilities, and individuals were visualized and analyzed.
LIMITATIONS: This study had limited sample diversity, and suggests further research is needed to fully understand the predictors of SUD relapse.
CONCLUSIONS: Interpretable machine learning not only performs well in predicting relapse but also helps clinical experts and healthcare workers to better understand each risk factor and each case. It also provides a new way of thinking about recognizing the risk factors that influence relapse.
METHODS: In this study, 5255 SUDs were collected by cluster-wide random sampling, including basic personal information, personal health, social relationships, drug abuse behaviors and characteristics, and other information collected to predict the recurrence rate. In segmenting the dataset for training and evaluating several widely used machine learning methods, the best performing eXtreme Gradient Boosting (XGboost) model is explained at different levels by SHapley Additive explanation (SHAP).
RESULTS: 1) The XGboost model achieved the best performance with an area under the receiver operating characteristic curve of 0.76 in the validation set. 2) Ranking of risk factors using SHAP found that drug use behaviors and characteristics were important features for the prediction of relapse behavior. 3) SHAP predicted and interpreted individual recurrence probabilities, and individuals were visualized and analyzed.
LIMITATIONS: This study had limited sample diversity, and suggests further research is needed to fully understand the predictors of SUD relapse.
CONCLUSIONS: Interpretable machine learning not only performs well in predicting relapse but also helps clinical experts and healthcare workers to better understand each risk factor and each case. It also provides a new way of thinking about recognizing the risk factors that influence relapse.