Skip to main content

Table 3 Simulation results part 2

From: Sparse and smooth canonical correlation analysis through rank-1 matrix approximation

  

θ x (rad)

θ y (rad)

θ x (rad)

θ y (rad)

θ x (rad)

θ y (rad)

 

Method

N=50

N=100

N=200

Scenario 4:

CCA

0.8125

0.9956

0.5603

0.6678

0.3390

0.4484

 

LS CCA

0.5275

0.7305

0.3553

0.4711

0.2412

0.3449

 

CCA LB

0.7603

0.9209

0.2785

0.5163

0.0149

0.3152

 

PMD

0.6111

0.8273

0.2031

0.4616

0.0397

0.3373

 

Algorithm 2

0.8829

0.9938

0.5288

0.6735

0.3295

0.4447

 

Algorithm 3

0.3990

0.6856

0.0173

0.3237

0.0001

0.3035

Scenario 5:

CCA

1.3798

1.3764

0.8879

0.8744

0.4700

0.4722

 

LS CCA

0.8538

0.8298

0.5231

0.5187

0.3373

0.3378

 

CCA LB

1.3681

1.3659

0.7264

0.7347

0.0478

0.0417

 

PMD

1.3972

1.3542

1.1316

1.0342

0.4082

0.3820

 

Algorithm 2

1.3627

1.3655

0.7413

0.8096

0.4407

0.4605

 

Algorithm 3

1.1185

1.0986

0.0275

0.0271

0.0001

0.0001

Scenario 6:

CCA

1.4853

1.4854

1.4624

1.4633

1.4249

1.4199

 

LS CCA

1.4589

1.4578

1.3797

1.3838

1.1954

1.1951

 

CCA LB

1.4862

1.4851

1.4684

1.4740

1.4830

1.4793

 

PMD

1.5244

1.5130

1.4985

1.4954

1.4553

1.4551

 

Algorithm 2

1.4794

1.4791

1.4512

1.4509

1.3869

1.3790

 

Algorithm 3

1.4633

1.4628

0.7775

0.7885

0.0220

0.0221