Development and validation comparison of multiple models for perioperative neurocognitive disorders during hip arthroplasty DOI Creative Commons
Gang Wang, Yi Xie, Xue‐Feng Bai

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Март 19, 2025

This study aims to develop optimal predictive models for perioperative neurocognitive disorders (PND) in hip arthroplasty patients, thereby advancing clinical practice. Data from all patients the MIMIC-IV database were utilized predict PND. With 62 variables, we applied multiple logistic regression, artificial neural network (ANN), Naive Bayes, support vector machine, and decision tree (XgBoost) algorithms forecast Feature analysis, receiver operating characteristic curve (ROC) calibration plotting, sensitivity, specificity, F-measure β = 1 (F1-score) assessments conducted on both training validation sets classifying models' effectiveness. Brier score Index of prediction accuracy (IPA) employed compare capabilities sets. Among 3,292 MIMIC database, 331 developed Five using different constructed. After thorough comparison validation, ANN model emerged as most effective model. Performance metrics set were: ROC: 0.954, Accuracy: 0.938, Precision: 0.758, F1-score: 0.657, Score: 0.048, IPA: 90.8%. On set, performed follows: 0.857, 0.903, 0.539, 0.432, 0.071, 71.4%. An online visualization tool was ( https://xyyy.pythonanywhere.com/ ).

Язык: Английский

Development and validation comparison of multiple models for perioperative neurocognitive disorders during hip arthroplasty DOI Creative Commons
Gang Wang, Yi Xie, Xue‐Feng Bai

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Март 19, 2025

This study aims to develop optimal predictive models for perioperative neurocognitive disorders (PND) in hip arthroplasty patients, thereby advancing clinical practice. Data from all patients the MIMIC-IV database were utilized predict PND. With 62 variables, we applied multiple logistic regression, artificial neural network (ANN), Naive Bayes, support vector machine, and decision tree (XgBoost) algorithms forecast Feature analysis, receiver operating characteristic curve (ROC) calibration plotting, sensitivity, specificity, F-measure β = 1 (F1-score) assessments conducted on both training validation sets classifying models' effectiveness. Brier score Index of prediction accuracy (IPA) employed compare capabilities sets. Among 3,292 MIMIC database, 331 developed Five using different constructed. After thorough comparison validation, ANN model emerged as most effective model. Performance metrics set were: ROC: 0.954, Accuracy: 0.938, Precision: 0.758, F1-score: 0.657, Score: 0.048, IPA: 90.8%. On set, performed follows: 0.857, 0.903, 0.539, 0.432, 0.071, 71.4%. An online visualization tool was ( https://xyyy.pythonanywhere.com/ ).

Язык: Английский

Процитировано

0