Table 3 Comparison of the Impact of Different Numbers of Clients on the Model

From: Intelligent diagnosis of gearbox in data heterogeneous environments based on federated supervised contrastive learning framework

Datasets

DDS

WTDS

Schemes

Metrics

Number of clients

4

6

8

4

6

8

IID

Accuracy

0.9936

0.9918

0.9921

0.9507

0.9499

0.9484

Specificity

0.9958

0.9947

0.9952

0.9574

0.9472

0.9469

Sensitivity

0.9764

0.9686

0.9675

0.8978

0.9711

0.9656

AUC

0.9995

0.9996

0.9993

0.9741

0.9728

0.9696

F1

0.9718

0.9633

0.9648

0.8472

0.8417

0.8351

Non-

IID- Class

Accuracy

0.9906

0.9876

0.9861

0.9322

0.9295

0.9122

Specificity

0.9937

0.9927

0.9862

0.9353

0.9333

0.9322

Sensitivity

0.9501

0.9300

0.9014

0.8989

0.8856

0.8722

AUC

0.9993

0.9980

0.9912

0.9631

0.9575

0.9495

F1

0.9516

0.9338

0.9153

0.7608

0.7466

0.7361

Non-

IID- Client

Accuracy

0.9593

0.9400

0.9325

0.9207

0.9169

0.9006

Specificity

0.9738

0.9646

0.9611

0.9493

0.9543

0.9544

Sensitivity

0.8205

0.7871

0.7450

0.8980

0.8914

0.8826

AUC

0.9860

0.9678

0.9695

0.9631

0.9575

0.9575

F1

0.8206

0.7814

0.7730

0.8470

0.7824

0.7746

Non-

IID- Domain

Accuracy

0.9892

0.9839

0.9703

0.9383

0.9310

0.9217

Specificity

0.9924

0.9900

0.9841

0.9356

0.9388

0.9371

Sensitivity

0.9630

0.9128

0.8836

0.9011

0.8689

0.8611

AUC

0.9991

0.9972

0.9935

0.9679

0.9646

0.9608

F1

0.9534

0.9135

0.8805

0.7845

0.7759

0.7645

Back to article page