Blind Federated Learning without initial model

Journal of Big Data

Table 8 Blind FL results (accuracy-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%	0.8636	0.8636	0.5000	0.5000
2	20%	26%	0.9130	0.9139	0.8333	0.7143
3	20%	35%	0.8696	0.8696	0.7778	0.7000
4	20%	13%	0.7826	0.8696	0.2500	0.4286
5	20%	26%	0.8261	0.9130	0.6000	0.7500
Avg.	–	–	0.8510	0.8858	0.5922	0.6186
1	19%	24%	0.9063	0.9063	0.6667	0.6667
2	26%	35%	0.9048	0.9048	0.8519	0.8519
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	0.8704	0.8636	0.8696
Avg.	–	–	0.8635	0.8672	0.8264	0.8276
1	5%	27%	1.0000	1.0000	1.0000	1.0000
2	4%	43%	0.8182	0.8182	1.0000	1.0000
3	42%	35%	0.9293	0.9394	0.9706	0.9375
4	20%	23%	0.8889	0.8889	0.7000	0.7000
5	29%	49%	0.9028	0.9028	0.9600	0.9583
Avg.	–	–	0.9078	0.9098	0.9261	0.9192
1	3%	20%	1.0000	1.0000	1.0000	1.0000
2	4%	36%	0.9231	0.9231	1.0000	1.0000
3	45%	40%	0.9515	0.9515	0.9737	0.9459
4	21%	17%	0.9512	0.9512	0.8889	0.8889
5	26%	49%	0.9219	0.9375	0.9615	1.0000
Avg.	–	–	0.9495	0.9527	0.9648	0.9670