Skip to main content

Table 1 The prediction results for four species via 10-fold cross-validation by SVM, ELM, MLP, CNN

From: Comparative analysis and prediction of nucleosome positioning using integrative feature representation and machine learning algorithms

Species

Method

K

ACC

Sn

Sp

MCC

AUC

H. sapiens

FCGR-SVM

1 + 2 + 4

0.8708

0.8980

0.8439

0.7432

0.9300

FCGR-ELM

2 + 4

0.8332

0.8773

0.7896

0.6695

0.8969

FCGR-MLP

3 + 4 + 5

0.8565

0.8768

0.8365

0.7144

0.9186

FCGR-CNN

4

0.8585

0.8746

0.8426

0.7185

0.9214

C. elegans

FCGR-SVM

1 + 2 + 4

0.8603

0.8948

0.8263

0.7229

0.9295

FCGR-ELM

1 + 2 + 4

0.8754

0.8944

0.8566

0.7515

0.9421

FCGR-MLP

3 + 4

0.8537

0.8613

0.8462

0.7092

0.9225

FCGR-CNN

4 + 5

0.8495

0.8839

0.8156

0.702

0.9181

D. melanogaster

FCGR-SVM

2 + 4

0.8113

0.7831

0.8400

0.6241

0.8791

FCGR-ELM

2

0.7910

0.7648

0.8175

0.5833

0.8595

FCGR-MLP

3 + 4 + 5

0.8117

0.8000

0.8235

0.6238

0.8848

FCGR-CNN

3 + 4 + 5

0.8108

0.8014

0.8204

0.6228

0.8854

S. cerevisiae

FCGR-SVM

4

1

1

1

1

1

FCGR-ELM

3 or 4

1

1

1

1

1

FCGR-MLP

4

1

1

1

1

1

FCGR-CNN

4

0.9997

1

0.9994

0.9995

1

  1. Best values are in bold