中医药运动康复疗效预测的深度学习模型及其应用研究OA

Objective:To enhance the predictive accuracy and clinical utility of rehabilitation outcomes by constructing an ensemble learning-based deep learning model to predict pain scores,functional assessment scale scores,and range of motion in patients undergoing Traditional Chinese Medicine(TCM)-based exercise rehabilitation.Methods:Clinical data,imaging data,TCM characteristic data,and weekly assessment data were collected from 717 patients undergoing exercise training.Data preprocessing included one-hot encoding for categorical variables and standardization with normalization for numerical variables.Feature extraction encompassed body mass index(BMI),initial pain scores,swelling and limited mobility status,MRI findings,baseline range of motion,electromyography(EMG)values,constitutional types,TCM syndromes,and herbal formula compositions.An ensemble learning model was constructed,and model performance was evaluated using 5-fold cross-validation.Results:The ensemble learning model demonstrated superior performance in predicting pain scores,functional assessment scale scores,and range of motion.Compared with individual models,the ensemble model achieved significant improvements in terms of mean absolute error(MAE),mean absolute percentage error(MAPE),and mean squared error(MSE),with particularly notable performance in predicting range of motion.Conclusion:The ensemble learning model effectively improves the predictive accuracy of rehabilitation outcomes in TCM-based exercise rehabilitation and demonstrates substantial clinical application value.