Table 6 Variable importance.

From: Traffic prediction in SDN for explainable QoS using deep learning approach

Variable

Relative importance

Scaled importance

Percentage

Dst

738.7526855

1.0

0.2245657

Src

394.7110901

0.5342939

0.1199841

Switch

278.0320129

0.3763533

0.0845161

dur_nsec

273.2421265

0.3698696

0.0830600

Protocol

272.5742493

0.3689655

0.0828570

flows

261.1378479

0.3534848

0.0793806

Pktcount

195.3378754

0.2644158

0.0593787

Packetins

152.1798248

0.2059956

0.0462596

Bytecount

136.2605286

0.1844467

0.0414204

tot_kbps

119.2972717

0.1614847

0.0362639

pktrate

81.5488281

0.1103872

0.0247892

port_no

77.5333252

0.1049517

0.0235685

tot_dur

41.8599777

0.0566630

0.0127246

Dur

32.4385338

0.0439099

0.0098607

rx_bytes

31.0710964

0.0420589

0.0094450

Pktperflow

25.5084591

0.0345291

0.0077541

tx_kbps

20.8022137

0.0281586

0.0063234

rx_kbps

20.0869160

0.0271903

0.0061060

tx_bytes

13.6969090

0.0185406

0.0041636

Pairflow

6.3214526

0.0085569

0.0019216