Table 3 Stability evaluation of feature subsets using the Kuncheva Index, and feature ranks using the complemented Canberra distance and Spearman’s rank correlation for the proposed Greedy-Graph (GG) and Brute-Graph (BG) methods, as well as the Classification-Based (CB), Phylogeny-Based (PB), and Leave-one-Variable-out (LV) methods

Iris

1.0000

1.0000

1.0000

1.0000

1.0000

1.0000

1.0000

1.0000

0.7977

1.0000

1.0000

1.0000

0.9191

Liver disorders

0.9642

0.9642

0.5840

1.0000

0.7676

0.9897

0.8670

1.0000

0.7267

0.9864

0.7818

1.0000

0.8184

Ecoli

0.9829

0.9829

1.0000

1.0000

0.7608

0.9848

0.8844

1.0000

0.9241

0.9920

0.9856

1.0000

0.8877

Breast tissue

0.9073

0.9073

0.6862

0.8802

0.6074

0.9414

0.7420

0.9350

0.8038

0.9725

0.8061

0.9637

0.7108

Glass

0.9637

0.9483

0.6966

0.9908

0.5323

0.9774

0.8250

0.9974

0.7712

0.9915

0.8350

0.9979

0.6327

Wine

0.9119

0.9100

0.7316

0.8726

0.2847

0.9324

0.8045

0.9118

0.6235

0.9827

0.8914

0.9672

0.2956

Lymphography

0.8838

0.8838

0.6469

0.9047

0.4852

0.9089

0.7782

0.9318

0.7251

0.9813

0.8344

0.9810

0.7482

Parkinson

0.9148

0.9024

0.4856

0.9081

0.1964

0.9632

0.7585

0.9371

0.6750

0.9883

0.5578

0.9893

0.4850

Ionosphere

0.9362

-

0.5952

0.9266

0.0639

0.9691

0.8158

0.9397

0.6554

0.9951

0.7825

0.9848

0.2718

Sonar

0.7814

-

0.3820

0.8340

0.2089

0.9259

0.7208

0.9226

0.7111

0.9879

0.4550

0.9756

0.5668

mean

0.9065

-

0.6289

0.9114

0.3761

0.9593

0.8196

0.9575

0.7414

0.9878

0.7930

0.9859

0.6336

p-value

<0.001

1.000

<0.001

<0.001

1.000

<0.001

<0.001

<0.001

<0.001