Skip to main content

Table 2 Comparison statistics on different constructed models using MLR, RBFNN and GEP techniques for PR, SL, STN, Vitri and QI of OHF and Pyrodwarf rootstocks during in vitro proliferation

From: Combining gene expression programming and genetic algorithm as a powerful hybrid modeling approach for pear rootstocks tissue culture media formulation

Rootstock

Measured parameter

Model

RMSE

MARE

MBE

R2

OHF

PR

MLR

1.82

0.534

0.030

0.583

 

RBFNN

0.69

0.138

− 0.014

0.957

 

GEP

0.48

0.128

− 0.015

0.992

SL

MLR

0.38

0.104

− 0.005

0.904

 

RBFNN

0.14

0.034

0.012

0.990

 

GEP

0.22

0.058

0.040

0.974

STN

MLR

10.77

0.523

0.184

0.647

 

RBFNN

6.95

0.211

0.118

0.882

 

GEP

4.21

0.138

− 0.096

0.961

Vitri

MLR

9.83

0.649

− 0.222

0.511

 

RBFNN

5.27

0.169

− 0.351

0.883

 

GEP

3.26

0.132

0.075

0.969

QI

MLR

0.87

0.269

− 0.019

0.668

 

RBFNN

0.52

0.155

− 0.028

0.891

 

GEP

0.45

0.160

− 0.018

0.943

Pyrodwarf

PR

MLR

3.58

1.160

− 1.572

0.315

 

RBFNN

1.49

0.379

0.087

0.908

 

GEP

0.38

0.071

− 0.003

0.994

SL

MLR

1.13

0.552

0.011

0.375

 

RBFNN

0.43

0.194

− 0.007

0.924

 

GEP

0.30

0.130

0.009

0.978

STN

MLR

6.48

0.217

− 0.438

0.855

 

RBFNN

5.42

0.158

− 0.456

0.906

 

GEP

4.22

0.126

0.045

0.954

Vitri

MLR

5.86

0.193

0.915

0.909

 

RBFNN

5.56

0.158

− 0.499

0.901

 

GEP

4.59

0.142

− 0.496

0.936

QI

MLR

0.47

0.171

− 0.063

0.902

 

RBFNN

0.43

0.163

0.032

0.923

 

GEP

0.41

0.160

− 0.018

0.944