Skip to main content

Table 7 The effect of predictive methods and growth stages on the MAE of ADG and F/G1,2

From: Predicting the growth performance of growing-finishing pigs based on net energy and digestible lysine intake using multiple regression and artificial neural networks models

Item

n

ADG

F/G

MR

ANN

P-value

MR

ANN

P-value

40-50 kg

16

87a,W ± 13

42b ± 7

< 0.01

0.21a,VW ± 0.03

0.12b,Y ± 0.01

< 0.01

50-60 kg

16

79W ± 11

78 ± 12

0.93

0.21a,V ± 0.12

0.12b,Y ± 0.05

< 0.01

60-70 kg

14

217a,X ± 29

84b ± 19

< 0.01

0.36a,VWX ± 0.24

0.17b ± 0.16

0.02

70-80 kg

12

191a,WX ± 27

81b ± 13

< 0.01

0.24VW ± 0.22

0.25 ± 0.17

0.9

80-90 kg

9

252a,XY ± 164

102b ± 65

0.02

0.47a,WXY ± 0.06

0.2b ± 0.03

< 0.01

90-100 kg

13

306a,XY ± 106

72b ± 40

< 0.01

0.77a,YZ ± 0.30

0.18b ± 0.11

< 0.01

100-110 kg

16

364a,YZ ± 117

81b ± 57

< 0.01

0.91a,Z ± 0.41

0.33b,Z ± 0.25

< 0.01

P-value

 

< 0.01

0.15

#

< 0.01

0.01

#

  1. MAE mean absolute error, ADG average daily gain, F/G feed conversion ratio.
  2. 1 Values are presented as means ± SEM. a-b in the same line means the MAE with different superscripts differ in predictive methods (P < 0.05). V-Z in the same column means the MAE with different superscripts differ in growth stages (P < 0.05). Pound sign means an interactive effect of methods and growth stages (P < 0.05).
  3. 2 The MAE were calculated by using predicted values and observed values in the validation data set (animal trial).