Journal of Cloud Computing

Advances, Systems and Applications

Table 5 The impact of different privacy protection components in SpFedRec on performance and efficiency. (Comm. Denote the communication overhead)

From: A split-federated learning and edge-cloud based efficient and privacy-preserving large-scale item recommendation model

Method	MovieLens			Adressa
Method	AUC	Training Time	Training Comm	AUC	Training Time	Training Comm
SpFedRec-DSSM	74.83	228 ms	1.35 MB	67.79	324 ms	2.18 MB
NO-OI	74.83	225 ms	1.35 MB	67.79	322 ms	2.18 MB
NO-SA	75.55	182 ms	0.91	68.10	279 ms	1.62 MB
SpFedRec-SENet	76.32	449 ms	1.73 MB	69.15	681 ms	4.40 MB
NO-OI	76.32	444 ms	1.73 MB	69.15	678 ms	4.40 MB
NO-SA	76.93	336 ms	1.02 MB	69.91	569 ms	3.26 MB

Back to article page