Combinatorial chemistry & high throughput screening · 2026
Development of An Explainable Machine Learning Model for Predicting Fatigue in Rheumatoid Arthritis.
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.
Ma Yucao, Zhang Yiyan, He Yunxi, Zhao Luming, Liu Chunping, Wang Hailong, Lv Liu
Аннотация
INTRODUCTION: Fatigue has been identified as one of the common symptoms among people who have rheumatoid arthritis (RA), which has a significant impact on their quality of life (QoL). Although fatigue has been identified as a symptom, few predictive models using clinical indicators have been developed for RA-related fatigue. This paper aims to create predictive models using machine learning (ML) algorithms.
METHODS: A retrospective analysis of clinical data was conducted on 271 patients with RA from two hospitals. The patients were randomly split into a training set and an external validation set. Feature selection on the training set was performed using three different algorithms: Least Absolute Shrinkage and Selection Operator (Lasso), Boruta, and Recursive Feature Elimination using a Random Forest (RF-RFE). Subsequently, multiple models using machine learning ML) techniques such as Support Vector Machine (SVM), XGBoost, LightGBM, Artificial Neural Network (ANN), K-Nearest Neighbors (KNN), and Random Forest (RF) were developed. Model calibration, Receiver Operating Curve (ROC) analyses, Decision Curve Analysis (DCA), and SHapley Additive exPlanations (SHAP) analyses were used to evaluate the models. Moreover, an overall evaluation using ten-fold cross-validation and an external evaluation using an independent dataset were undertaken.
RESULTS: The final dataset comprised 190 patients in the training group and 81 in the external validation set, with no significant differences in demographic features such as age and sex (P > 0.05). The six important predictors identified using a variety of feature selectors were CCP, ESR, lymphocyte count, MDGA, VAS for pain, and TG. In external validation, ML model classifiers such as SVM, XGBoost, and Random Forest performed better at prediction than logistic regression.
DISCUSSION: This study demonstrates the potential of ML models to capture nonlinear relationships between clinical variables when predicting fatigue in RA patients. However, the modest performance metrics suggest that these findings are preliminary.
CONCLUSION: A fatigue-prediction model using routine clinical parameters was developed using ML techniques. The SVM model showed potential in identifying RA patients at risk of fatigue. However, the modest performance metrics indicate the preliminary nature of these results, necessitating validation through larger, multicenter studies.
METHODS: A retrospective analysis of clinical data was conducted on 271 patients with RA from two hospitals. The patients were randomly split into a training set and an external validation set. Feature selection on the training set was performed using three different algorithms: Least Absolute Shrinkage and Selection Operator (Lasso), Boruta, and Recursive Feature Elimination using a Random Forest (RF-RFE). Subsequently, multiple models using machine learning ML) techniques such as Support Vector Machine (SVM), XGBoost, LightGBM, Artificial Neural Network (ANN), K-Nearest Neighbors (KNN), and Random Forest (RF) were developed. Model calibration, Receiver Operating Curve (ROC) analyses, Decision Curve Analysis (DCA), and SHapley Additive exPlanations (SHAP) analyses were used to evaluate the models. Moreover, an overall evaluation using ten-fold cross-validation and an external evaluation using an independent dataset were undertaken.
RESULTS: The final dataset comprised 190 patients in the training group and 81 in the external validation set, with no significant differences in demographic features such as age and sex (P > 0.05). The six important predictors identified using a variety of feature selectors were CCP, ESR, lymphocyte count, MDGA, VAS for pain, and TG. In external validation, ML model classifiers such as SVM, XGBoost, and Random Forest performed better at prediction than logistic regression.
DISCUSSION: This study demonstrates the potential of ML models to capture nonlinear relationships between clinical variables when predicting fatigue in RA patients. However, the modest performance metrics suggest that these findings are preliminary.
CONCLUSION: A fatigue-prediction model using routine clinical parameters was developed using ML techniques. The SVM model showed potential in identifying RA patients at risk of fatigue. However, the modest performance metrics indicate the preliminary nature of these results, necessitating validation through larger, multicenter studies.