Table 13 Statistical significance test results on the MARS dataset.

From: Adaptive course recommendation using federated learning and graph convolutional networks in IoT-enhanced e-learning

Model

Precision@10 (mean ± std)

NDCG@10 (mean ± std)

MAE (mean ± std)

p-value vs. proposed

Proposed (FL + GCN + DistilBERT + IoT)

0.885 ± 0.006

0.880 ± 0.005

0.215 ± 0.004

–

LightGCN–PKA

0.875 ± 0.008

0.870 ± 0.007

0.275 ± 0.006

< 0.01

FGC

0.849 ± 0.010

0.852 ± 0.009

0.249 ± 0.007

< 0.01

Transformer-based FL

0.855 ± 0.007

0.840 ± 0.008

0.255 ± 0.006

< 0.05

Back to article page