基于生物电阻抗技术的肉种鸡母鸡腹部脂肪含量预测研究  

Prediction of Abdominal Fat Content in Broiler Breeder Hens Based on Bioelectrical Impedance Technology

作　　者：刘健 吴锦辉 祝鹏飞 陈威 滕光辉 LIU Jian;WU Jinhui;ZHU Pengfei;CHEN Wei;TENG Guanghui(College of Water Resources and Civil Engineering,China Agricultural University,Beijing 100083;Key Laboratory of Agricultural Engineering in Structure and Environment,Ministry of Agriculture and Rural Affairs,Beijing 100083)

机构地区：[1]中国农业大学水利与土木工程学院,北京100083 [2]农业农村部设施农业工程重点实验室,北京100083

出　　处：《中国家禽》2025年第4期89-95,共7页China Poultry

基　　金：国家重点研发计划项目(2023YFD2000800);山东省重点研发计划(乡村振兴科技创新提振行动计划)(2022TZXD0015)。

摘　　要：为探究肉种鸡母鸡特征参数(阻抗幅值、相位角、体重、体斜长、胸围、周龄)与腹脂重之间的量化关系,实现腹部脂肪含量的快速检测,研究以70只不同周龄的Cobb500肉种鸡母鸡为研究对象,采用皮尔逊相关系数为评价指标,确定肉种鸡母鸡阻抗的最优测量点位,并以该测量点位下的阻抗幅值、相位角和其余特征参数(体重、体斜长、胸围、周龄)为特征输入,以腹脂重作为预测指标,分别采用多元线性回归模型和4种机器学习回归模型(随机森林回归模型、支持向量回归模型、BP神经网络、极限梯度提升模型)进行建模分析。结果显示:相较于多元线性回归模型,机器学习回归模型的预测准确性普遍较高,其中极限梯度提升模型表现出最佳性能,在测试集上的决定系数(R^(2))、均方误差(MSE)、平均绝对误差(MAE)、解释方差分(EVS)分别为0.908 3、10.84、168.87、0.908 8。研究表明,极限梯度提升模型在预测肉种鸡母鸡腹脂重方面表现出优越性能,适用于肉种鸡母鸡养殖过程中腹部脂肪含量的预测,结果也为肉种鸡福利化养殖提供了技术支撑和科学依据。In order to explore the quantitative relationship between the characteristic parameters(impedance amplitude,phase angle,body weight,body slope length,body girth and week age)and abdominal fat weight of broiler breeder hen,70 Cobb500 broiler breeder hens with different weeks of age were taken as the objects.The pearson correlation coefficient was used as evaluation index,the optimal measuring point of impedance of broiler breeder hen was determined,the impedance amplitude,phase angle and other characteristic parameters(body weight,body slope length,body girth,week age)at this measuring point were used as characteristic inputs,abdominal fat weight was used as a predictive index,multiple linear regression model and four kinds of machine learning regression models(random forest regression model,support vector regression model,BP neural network and extreme gradient lifting model)were used for modeling and analysis respectively.The results showed that the prediction accu-racy of machine learning regression model was generally higher than that of multiple linear regression model,and the ultimate gra-dient lifting model showed the best performance,and the determination coefficient(R^(2)),mean square error(MSE),mean absolute error(MAE)and interpretive difference(EVS)on the test set were 0.9083,10.84,168.87 and 0.9088,respectively.The results indicated that the limit gradient lifting model was superior in predicting the abdominal fat weight of broiler breeder hens,which al-so provided some technical support and scientific basis for the welfare breeding of broiler breeders.

关 键 词：肉种鸡 母鸡 生物电阻抗技术 机器学习 腹部脂肪 特征参数 

分 类 号：S831.4[农业科学—畜牧学]