| 摘 要: 为预测奶牛酮病,建立基于改进XGBoost的奶牛酮病预测模型,提高该疾病的预测效率。选取甘肃某牧场的奶牛生产性能测定DHI数据进行分析,将脂蛋比作为奶牛酮病的预警指标,建立了基于IGJO-XGBoost回归预测模型来预测奶牛酮病。同时构建不同的奶牛酮病回归预测模型进行对比实验,评估模型的性能。实验结果表明,改进后的XGBoost奶牛酮病回归预测模型的各项评价指标均优于其他模型,在奶牛酮病预测方面具有较好的预测结果,可以为奶牛酮病的预防提供参考。 |
| 关键词: 奶牛酮病 回归预测 金豺优化算法 XGBoost |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 甘肃省高等学校创新基金项目(2022B-107,2019A-056);甘肃农业大学青年导师基金项目(GAU-QDFC-2019-02);甘肃省自然科学基金项目
(20JR10RA510,1506RJZA007) |
|
| Research on a Predictive Model for Dairy Cow Ketosis Based on Improved XGBoos |
|
CHEN Shifei, DAI Yongqiang
|
(College of In formation Science and Technology, Gansu Agricultural University, Lanzhou 730070, China)
csf526408@163.com; dyq@gsau.edu.cn
|
| Abstract: To predict dairy cow ketosis and enhance prediction efficiency, this study establishes an improved XGBoost-based prediction model. Using Dairy Herd Improvement (DHI) data from a dairy farm in Gansu Province, the
fat-to-protein ratio was selected as an early warning indicator for ketosis. An IGJO-enhanced XGBoost regression prediction model was developed to forecast ketosis incidence. Comparative experiments with other regression prediction models were conducted to evaluate performance. Results demonstrate that the improved XGBoost model outperforms others across all evaluation metrics, exhibiting superior predictive capability for dairy cow ketosis. This model provides valuable insights for ketosis prevention strategies. |
| Keywords: dairy cow ketosis regression prediction Golden Jackal Optimization (GJO) XGBoost |
