Table 2 Details of experimental results based on NMI scores for various dimension reduction algorithms, including the DR-A, PCA, ZIFA, scVI, SAUCIE, t-SNE, and UMAP methods. We carried out the experiments using the Rosenberg-156 k, Zheng-73 k, Zheng-68 k, Macosko-44 k, and Zeisel-3 k datasets. These dimension reduction algorithms were investigated with (a) 2 latent dimensions (K = 2), (b) 10 latent dimensions (K = 10), and (c) 20 latent dimensions (K = 20)

 DR-A

0.5573

0.8457

0.5931

0.4936

0.7263

 PCA

0.2523

0.3396

0.2538

0.2984

0.4721

 ZIFA

0.3049

0.3794

0.2810

0.3120

0.4250

 scVI

0.5199

0.8261

0.5417

0.4599

0.7006

 SAUCIE

0.4046

0.4304

0.2749

0.2707

0.4622

 t-SNE

0.4343

0.6562

0.4081

0.4091

0.7103

 UMAP

0.5591

0.6507

0.4377

0.4184

0.7214

 DR-A

0.5850

0.8503

0.5756

0.5156

0.7893

 PCA

0.3276

0.5612

0.3877

0.4243

0.5559

 ZIFA

0.5074

0.8354

0.5152

0.4785

0.7807

 scVI

0.5821

0.8060

0.5571

0.5155

0.7606

 SAUCIE

0.4773

0.4209

0.3147

0.2874

0.5110

 t-SNE

N/A

N/A

N/A

N/A

N/A

 UMAP

0.5735

0.6911

0.4393

0.4129

0.7413

 DR-A

0.5842

0.8002

0.5888

0.5176

0.7639

 PCA

0.3761

0.5623

0.3874

0.4306

0.5561

 ZIFA

N/A

N/A

N/A

N/A

0.7114

 scVI

0.5831

0.7976

0.5691

0.5105

0.7419

 SAUCIE

0.4740

0.4254

0.2952

0.2775

0.4808

 t-SNE

N/A

N/A

N/A

N/A

N/A

 UMAP

0.5656

0.6906

0.4413

0.4177

0.7419