A machine learning-based prediction model for postoperative delirium in cardiac valve surgery using electronic health records

Table 3 ECE in the test dataset and the 5 specific population subsets

ECE	Test dataset (n = 102)	Age ≥ 65 (n = 131)	BMI ≤ 18.5 (n = 64)	BMI>28 (n = 29)	Cerebral infarction (n = 33)	Coronary heart disease (n = 34)
Algorithms
Random Forest Classifier	0.10	0.13	0.16	0.14	0.19	0.21
Gradient Boosting Decision Tree	0.05	0.02	0.01	0.07	0.05	0.08
Support Vector Machine Classifier	0.15	0.28	0.30	0.29	0.32	0.21
Logistic Regression	0.06	0.06	0.11	0.13	0.13	0.04
Gaussian Naive Bayes	0.08	0.14	0.13	0.14	0.08	0.21
K-nearest Neighbors Classifier	0.14	0.06	0.04	0.15	0.21	0.24
Perceptron	0.23	0.20	0.17	0.21	0.24	0.14

The test dataset includes all features (n = 102)
Specific population subset 1: Patients aged ≥ 65 (n = 131); Specific population subset 2: Patients with BMI ≤ 18.5 (n = 64); Specific population subset 3: Patients with BMI > 28 (n = 29); Specific population subset 4: Patients with a history of cerebral infarction (n = 33); Specific population subset 5: Patients with a history of coronary heart disease (n = 34)

ISSN: 1471-2261