Raw performance metrics for the MLP model.

The table presents the raw output from the bootstrap analysis, showing the mean, lower, and upper confidence interval bounds for the AUC of local and federated models, and the performance gain (ΔAUC). It provides a hospital-level breakdown of the mean AUC and 95% CI for the local and federated MLP models. The table also includes the mean ΔAUC and its confidence interval, allowing for an assessment of the performance gain achieved by the federated neural network. (XLSX)