Study on the influence of different production factors on PSY and its correlation.

Background: Finding out the key reproductive performance factors, affecting piglets weaned per sow per year (PSY) can improve the production efficiency and profitability of pig farms. The objective was to understand the actual distribution of different production factors and PSY of breeding pig farms, analyze the correlation to find the main production factors affecting PSY, and formulating a Production Efficiency Improvement Plan in practice. Data included 603 breeding pig farms from September 28, 2020 to September 26, 2021. Regression analysis was used to evaluate the relationship between PSY and key production factors, and the characteristics of total pig farms versus high performance (HP) pig farms (the production performance was in the top 10%) or top 5% pig farms were compared. Spearman's rank correlation coefficient was used to analyze the correlation between production factors and find the factors related to PSY. Non-linear support vector regression (NL-SVR) was used to analyze the personalized PSY improvement through a various change of the four key factors.

Results: The median distribution of 15 production factors and PSY in total pig farms were different from those of HP farms. All of data were distributed nonlinearly. Mating rate within 7 days after weaning (MR7DW), farrowing rate (FR), number of piglets born alive per litter (PBAL) and number of weaned piglets per litter (WPL) were moderately correlated with PSY, and the correlation coefficients were 0.5058, 0.4427, 0.3929 and 0.3839, respectively. When the four factors in NL-SVR changed in medium (0.5 piglet or 5%) or high level (1.0 piglet or 10%), PSY can be increased by more than 0.5.

Conclusion: NL-SVR model can be used to analyze the impact of changes in key production factors on PSY. By taking measures to improve MR7DW, FR, PBAL and WPL, it may effectively improve the current PSY and fully develop the reproductive potential of sows.