Table 3 Performance comparison of models.

From: A hybrid federated learning framework with generative AI for privacy-preserving and sustainable security in IOT-enabled smart environments

Model scenario

Test accuracy (%)

F1-score

Precision

Recall

Centralized baseline

98.5%

0.98

0.99

0.98

Basic federated learning (FL)

95.2%

0.95

0.94

0.96

FL with differential privacy

94.1%

0.94

0.93

0.95

HFL-GAI (proposed)

96.8%

0.97

0.97

0.97

Back to article page