Table 3 Comparison of performance of clustering methods on eleven real UCI datasets

Normalised Mutual Information (NMI)(Number of clusters fixed to ground truth: c=c^∗)

Iris (c^∗=3)

0.7582

0.8997

0.7221

0.7578

0.7665

0.7980

Glass (c^∗=6)

0.3163

0.4033

0.1548

0.2899

0.3175

0.3932

Wine (c^∗=3)

0.8759

0.9120

0.6144

0.9276

0.9276

0.8347

WBDC (c^∗=2)

0.5546

0.5275

0.0102

0.5434

0.5565

0.7231

Control chart (c^∗=6)

0.6907

0.5348

0.7070

0.7924

0.7891

0.8489

Parkinson (c^∗=2)

0.0970

0.0387

0.0519

0.2189

0.2189

0.1334

Vertebral (c^∗=3)

0.4120

0.5559

0.1280

0.3969

0.3901

0.5971

Breast tissue (c^∗=6)

0.5446

0.5487

0.4745

0.5481

0.5366

0.5291

Seeds (c^∗=3)

0.7279

0.8111

0.7010

0.6881

0.6881

0.7142

Image Seg. (c^∗=7)

0.5840

0.5697

0.0359

0.5929

0.6466

0.5816

Yeast (c^∗=10)

0.2955

0.2071

0.1846

0.2794

0.2801

0.2909

Average

0.5324

0.5462

0.3440

0.5487

0.5561

0.5858

Adjusted Rand Index (ARI)(Number of clusters fixed to ground truth: c=c^∗)

Iris (c^∗=3)

0.7302

0.9039

0.6423

0.7430

0.7570

0.7455

Glass (c^∗=6)

0.1722

0.2649

0.0350

0.1104

0.1483

0.2380

Wine (c^∗=3)

0.8975

0.9325

0.5771

0.9471

0.9471

0.8471

WBDC (c^∗=2)

0.6707

0.6436

0.0048

0.6661

0.6777

0.8244

Control chart (c^∗=6)

0.5207

0.2699

0.5374

0.6835

0.6803

0.7056

Parkinson (c^∗=2)

-0.0978

-0.0502

-0.0710

0.1608

0.1608

0.2659

Vertebral (c^∗=3)

0.3147

0.5256

-0.0327

0.2986

0.2888

0.6096

Breast tissue (c^∗=6)

0.2876

0.2872

0.1895

0.3983

0.3806

0.3659

Seeds (c^∗=3)

0.7733

0.8167

0.6863

0.7189

0.7189

0.7432

Image Seg. (c^∗=7)

0.4605

0.4522

0.0011

0.3360

0.4849

0.4150

Yeast (c^∗=10)

0.1658

0.0972

0.1083

0.1483

0.1481

0.1746

Average

0.4450

0.4676

0.2435

0.4737

0.4902

0.5395