Blind Federated Learning without initial model

Journal of Big Data

Table 13 Blended Blind FL results (Precision-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%	0.8182	0.9091	0.4286	0.6000
2	20%	26%	0.9130	0.9130	0.8333	0.8333
3	20%	35%	0.8261	0.8696	0.7000	0.7778
4	20%	13%	0.9130	0.9130	0.6000	0.6000
5	20%	26%	0.9130	1.0000	0.7500	1.0000
Avg.	–	–	0.8767	0.9209	0.6624	0.7622
1	19%	24%	0.9545	0.9773	0.9000	0.8889
2	26%	35%	0.9000	0.9000	0.8947	0.8889
3	16%	42%	0.8333	0.8611	1.0000	1.0000
4	15%	20%	0.8529	0.8529	0.5833	0.5833
5	24%	47%	0.8727	0.8727	0.9524	0.9500
Avg.	–	–	0.8827	0.8928	0.8661	0.8622
1	5%	27%	0.7500	0.7500	0.5000	0.5000
2	4%	43%	0.8000	0.8000	1.0000	1.0000
3	42%	35%	0.9063	0.9167	0.8788	0.9063
4	20%	23%	0.9778	0.9778	0.9091	0.9091
5	29%	49%	0.7727	0.7727	0.9500	0.8261
Avg.	–	–	0.8414	0.8434	0.8476	0.8283
1	3%	20%	0.8333	0.8333	0.5000	0.5000
2	4%	36%	0.9000	0.9000	1.0000	1.0000
3	45%	40%	0.9519	0.9519	0.9512	0.9487
4	21%	17%	0.8333	0.8750	0.5000	0.5833
5	26%	49%	0.8361	0.8689	0.9565	1.0000
Avg.	–	–	0.8709	0.8858	0.7815	0.8064