Blind Federated Learning without initial model

Journal of Big Data

Table 12 Blind FL results (Precision-based weights, slope 5, sigmoid)—Breast Cancer dataset

Agent	Size	% 1s	Accuracy	Accuracy	Precision	Precision
Agent	Size	% 1s	Pre-FL	Post-FL	Pre-FL	Post-FL
1	20%	14%	0.5909	0.9091	0.2500	0.6000
2	20%	26%	0.9130	0.9130	0.8333	0.8333
3	20%	35%	0.8696	0.8696	0.7778	0.7778
4	20%	13%	0.7826	0.8696	0.3750	0.5000
5	20%	26%	0.9565	1.0000	0.8571	1.0000
Avg.	–	–	0.8225	0.9123	0.6187	0.7422
1	19%	24%	0.8750	0.8750	0.6000	0.6000
2	26%	35%	0.9048	0.9206	0.8800	0.8846
3	16%	42%	0.7317	0.7317	1.0000	1.0000
4	15%	20%	0.9231	0.9231	0.7500	0.7500
5	24%	47%	0.8519	0.8889	0.8636	0.9091
Avg.	–	–	0.8573	0.8679	0.8187	0.8287
1	5%	27%	0.9000	0.9000	0.7500	0.7500
2	4%	43%	0.9286	0.9286	0.8333	0.8333
3	42%	35%	0.9035	0.9035	0.9024	0.8810
4	20%	23%	0.8974	0.9231	0.7500	0.7500
5	29%	49%	0.9423	0.9423	1.0000	1.0000
Avg.	–	–	0.9144	0.9195	0.8472	0.8429
1	3%	20%	0.8333	0.8333	0.3333	0.3333
2	4%	36%	0.8571	0.9286	0.8000	0.8333
3	45%	40%	0.9107	0.9107	1.0000	0.9697
4	21%	17%	0.8857	0.8857	0.7143	0.7143
5	26%	49%	0.8621	0.8793	0.9583	0.9600
Avg.	–	–	0.8698	0.8875	0.7612	0.7621