Table 2 Overall performance comparison, where bold indicates the best performance for each row, underlined indicates the best baseline performance, and %Improv. represents the performance improvement of GR-MC compared to the best baseline.

From: A graph neural network recommendation algorithm based on multi-scale attention and contrastive learning

Methods

Movielens-100K

Movielens-1M

Gowalla

Yelp2018

Amazon-book

 

R@20

N@20

R@20

N@20

R@20

N@20

R@20

N@20

R@20

N@20

MF

0.0696

0.0467

0.1319

0.1327

0.1291

0.1109

0.0433

0.0354

0.0250

0.0196

NeuMF

0.0719

0.0487

0.1389

0.1395

0.1399

0.1212

0.0451

0.0363

0.0258

0.0200

GC-MC

0.0813

0.0443

0.1554

0.1557

0.1395

0.1204

0.0462

0.0379

0.0288

0.0224

NGCF

0.1433

0.1817

0.2437

0.2330

0.1569

0.1327

0.0579

0.0477

0.0337

0.0261

LightGCN

0.1483

0.1933

0.2501

0.2412

0.1830

0.1554

0.0649

0.0530

0.0411

0.0315

GTN

0.1472

0.1933

0.2654

0.2543

0.1906

0.1173

0.0637

0.0523

0.0465

0.0365

NGAT4Rec

0.1496

0.1988

0.2823

0.2659

0.2009

0.1185

0.0675

0.0554

0.0457

0.0358

SGL

0.1502

0.1998

0.2851

0.2662

0.2018

0.1196

0.0675

0.0555

0.0478

0.0379

GR-MC

0.1646

0.2095

0.2971

0.3380

0.2081

0.1640

0.0731

0.0627

0.0596

0.0462

%Improv.

9.59%

4.85%

4.21%

26.97%

3.12%

5.53%

8.30%

12.97%

24.69%

21.90%

Back to article page