Skip to main content
. 2022 Sep 5;19(4):20220032. doi: 10.1515/jib-2022-0032

Table 1:

Results of the centralized and FL analysis, showing the mean accuracy, ROC AUC, sensitivity, specificity and number of patient samples for Institution 1 and 2 individually, the FL model and the centralized ML model.

Metric Institution 1 Institution 2 Centralized Federated
Accuracy 68.12% 64.71% 67.65% 67.65%
ROC AUC 75.22% 71.14% 75.52% 75.09%
Sensitivity 64.86% 66.67% 65.52% 66.67%
Specificity 71.61% 63.64% 69.70% 68.57%
Samples 477 202 680 477 + 202