Skip to main content

Table 17 Blind FL results (constant weights)—adult dataset

From: Blind Federated Learning without initial model

Agent

Size

% 1s

Accuracy

Accuracy

Precision

Precision

Pre-FL

Post-FL

Pre-FL

Post-FL

1

20%

72%

0.9444

0.9444

0.9615

0.9259

2

20%

75%

0.8055

0.8611

0.8846

0.9230

3

20%

75%

0.8611

0.9166

0.8928

0.9600

4

20%

67%

0.8888

0.9166

0.9545

0.9565

5

20%

72%

0.9444

0.9722

0.9615

1.0000

Avg.

0.8889

0.9222

0.9310

0.9531

1

13%

63%

0.5000

0.5000

1.0000

1.0000

2

30%

80%

0.3261

0.3478

1.0000

1.0000

3

21%

87%

0.2821

0.3077

1.0000

1.0000

4

11%

61%

0.6047

0.6279

1.0000

1.0000

5

24%

81%

0.3824

0.3922

0.9615

0.9412

Avg.

0.4190

0.4351

0.9923

0.9882

1

4%

75%

0.3125

0.3125

1.0000

0.0000

2

5%

68%

0.3684

0.3684

0.6666

0.6666

3

45%

77%

0.3496

0.3619

0.8571

0.8260

4

13%

51%

0.5869

0.5869

1.0000

1.0000

5

32%

85%

0.2649

0.2820

0.8823

0.8125

Avg.

0.3765

0.3823

0.8812

0.6610

1

1%

25%

0.3333

0.3333

0.0000

0.0000

2

5%

72%

0.3500

0.3500

1.0000

0.0000

3

45%

78%

0.2546

0.2732

0.8750

0.8750

4

14%

53%

0.4693

0.4693

0.0000

0.0000

5

35%

89%

0.1825

0.1904

1.0000

1.0000

Avg.

0.3179

0.3232

0.5750

0.3750

  1. Bold values indicate the best results