Cross-Site Validation

How it works

Benefits

Evidence quality

Highlighted claims

• The protocol uses nested cross-validation to evaluate the complete modelling process. — Development and cross-site validation of machine-learning models for diagnosis and prognosis of stable angina with and without obstructive coronary artery disease: a study protocol
• Models are trained on four sites and validated on the remaining site, repeated until every site has been held out once. — Development and cross-site validation of machine-learning models for diagnosis and prognosis of stable angina with and without obstructive coronary artery disease: a study protocol
• The inner loop handles feature selection, hyperparameter tuning, threshold tuning, and model selection. — Development and cross-site validation of machine-learning models for diagnosis and prognosis of stable angina with and without obstructive coronary artery disease: a study protocol
• The outer loop estimates performance on held-out external sites using leave-one-site-out cross-validation. — Development and cross-site validation of machine-learning models for diagnosis and prognosis of stable angina with and without obstructive coronary artery disease: a study protocol
• Nested cross-validation with leave-one-site-out validation is presented as a strength because it provides repeated external validation and less biased performance estimates. — Development and cross-site validation of machine-learning models for diagnosis and prognosis of stable angina with and without obstructive coronary artery disease: a study protocol