Blind Federated Learning without initial model

Journal of Big Data

Table 9 Blended Blind FL results (accuracy-based weights)—Breast Cancer dataset

Agent	Size	% 1s	Accuracy	Accuracy	Precision	Precision
Agent	Size	% 1s	Pre-FL	Post-FL	Pre-FL	Post-FL
1	20%	14%	0.7273	0.9091	0.3333	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%	0.9130	0.9130	0.6000	0.5000
5	20%	26%	0.8696	1.0000	0.6666	1.0000
Avg	–	–	0.8585	0.9209	0.6422	0.7422
1	5%	27%	0.9545	0.9773	0.9000	1.0000
2	4%	43%	0.9167	0.9167	0.9000	0.8947
3	42%	35%	0.8611	0.8611	1.0000	1.0000
4	20%	23%	0.9118	0.9118	0.7000	0.5833
5	29%	49%	0.8909	0.8909	0.9545	0.9545
Avg.	–	–	0.9070	0.9115	0.8909	0.8865
1	19%	24%	0.6667	0.7500	0.4000	0.5000
2	26%	35%	0.8000	0.8000	1.0000	1.0000
3	16%	42%	0.9063	0.9167	0.8571	0.8824
4	15%	20%	0.9778	0.9778	0.9091	0.9091
5	24%	47%	0.7576	0.7576	0.9048	0.9048
Avg.	–	–	0.8217	0.8404	0.8142	0.8392
1	3%	20%	0.8333	1.0000	0.5000	1.0000
2	4%	36%	0.9000	0.9000	1.0000	1.0000
3	45%	40%	0.9615	0.9615	0.9750	0.9487
4	21%	17%	0.8333	0.8958	0.5000	0.6364
5	26%	49%	0.8525	0.8525	1.0000	1.0000
Avg.	–	–	0.8761	0.9220	0.7950	0.9170