Skip to main content

Table 2 Personalized bottleneck calculated by gradient boosting regressor model*

From: Machine learning based personalized promotion strategy of piglets weaned per sow per year in large-scale pig farms

Factors

Number of pig farms

The proportion of pig farms

0.5 SD

Average

Median

Max

Min

Production days

101

34.7%

163.60

1.41

1.44

2.92

0.29

Number of weaned piglets per litter

60

20.6%

0.46

1.14

1.15

1.97

0.22

Mating rate within 7 days after weaning

45

15.5%

6.99%

1.63

1.58

3.31

0.50

Farrowing rate

24

8.2%

4.06%

1.01

1.07

1.76

0.28

Designed stock

17

5.8%

299.90

1.33

1.29

2.37

0.61

Number of piglets born alive per litter

15

5.2%

0.40

1.04

1.05

1.45

0.61

Weaning to breeding interval

11

3.8%

4.39

1.84

1.82

2.57

1.16

Non-productive days

4

1.4%

14.15

1.35

1.06

2.93

0.35

Actual stock

3

1.0%

300.01

1.56

1.50

2.74

0.42

Full load rate

3

1.0%

4.58%

0.32

0.32

0.36

0.28

Birth weight of piglets

2

0.7%

0.06

0.67

0.67

1.01

0.33

Mummified piglet rate

2

0.7%

0.36%

0.27

0.27

0.38

0.15

Total number of piglets per litter

2

0.7%

0.40

0.80

0.80

1.21

0.38

21-day adjusted weight of piglets

1

0.3%

0.18

0.34

0.34

0.34

0.34

Stillbirth rate

1

0.3%

1.01%

0.28

0.28

0.28

0.28

  1. For PSY improvement of 291 large-scale pig farms, each factor was increased by the absolute value of 0.5 standard deviation
  2. SD = standard deviation 
  3. *Return-service rate and farm type did not affect the PSY improvement by gradient boosting regressor