运用线性回归模型预测不同月龄肉用褐牛体重

A linear regression model was used to predict the weight of brown cattle for meat at different ages

中文关键词: 新疆褐牛  回归模型  相关性  估测体重

英文关键词: Xinjiang brown cattle  Regression model  Relevance  Weight estimation

中文摘要:

      【目的】体尺体重是衡量牛生长发育的重要指标，对饲养管理与选种选育具有重要意义。为了构建新疆褐牛公母牛体尺与体重回归预测模型。【方法】在本研究中选取了不同月龄阶段健康状况良好、品种特征明显的新疆褐牛公母牛共2045头，对不同年龄和性别下的新疆褐牛进行体尺及体重指标测定，并将数据进行了分析，通过对比13个体尺指标和仅用相关系数最高的4个体尺指标与体重的关系，构建了相应的体尺与体重回归模型。【结果】通过对比 建立的模型，我们研究发现胸围和腹围是影响体重的主要因素。特别是对于0-6月龄的，使用13个体尺指标建立的线性回归模型具有最高的拟合度分别为0-6月龄的公牛和母牛，为 0.918和0.942，0-6月龄公牛线性回归方程：Y=-165.185+0.028X1-0.088X2-0.035X3+0.239X4+1.668X5+0.109X6+1.053X7-0.439X8+1.068X9+1.796X10+1.573X11-0.802X12-0.258X13 (R2=0.918)，0-6月龄母牛线性回归方程：Y=-182.399+0.97X1-0.359X2+0.579X3-0.097X4-0.683X5+0.025X6-0.364X7+0.502X8+0.329X9+1.374X10+1.786X11-0.083X12+0.097X13(R2=0.942)；使用4个体尺指标建立的线性回归模型具有最高的拟合度分别为0.882和0.933，0-6月龄公牛线性回归方程：Y=-162.511+0.555X2+1.218X9+0.986X11-0.127X12(R2=0.882)，0-6月龄母牛线性回归方程：Y=-175.941+0.436X2+0.449X9+2.073X11-0.213X12 (R2=0.933)。不同性别的牛群在各年龄段生长发育速度不同，除13-18月龄的公牛外，在此年龄段发育迅速，变化大，动态明显。导致体尺与体重的回归模型拟合效果不是很好，但是在其他不同年龄段和不同性别的模型展现了良好的拟合效果。【结论】新疆褐牛的体尺与体重之间存在显著相关性，建立了体尺预测体重的线性回归公式，且0-6月龄、7-12月龄以及30月龄以上的牛群的模型通过成对样本t检验得出预测体重与实测体重呈现不显著，可以用来预测牛群的体重。这对新疆褐牛不同年龄段不同性别的生长发育检测，体重预测有了一定的生产实践。

英文摘要:

      【 Objective 】 Body size and weight is an important index to measure the growth and development of cattle, which is of great significance to feeding management and seed selection. In order to construct the regression prediction model of body size and weight of Xinjiang brown cattle and cows. 【 Method 】 In this study, a total of 2045 Xinjiang brown cows with good health and obvious breed characteristics were selected. The body measurements and weight indexes of Xinjiang brown cows of different ages and sexes were measured, and the data were analyzed. By comparing the relationship between 13 body measurements and only 4 body measurements with the highest correlation coefficient, the corresponding regression model between body measurements and weight was constructed. 【 Results 】 By comparing the established models, we found that chest circumference and abdominal circumference are the main factors affecting body weight. Especially for bulls and cows aged 0-6 months, the linear regression models established by using 13 body measurements have the highest fitting degrees of 0.918 and 0.942 respectively. The linear regression equation of bulls aged 0-6 months is:Y=-165.185+0.028X1-0.088X2-0.035X3+0.239X4+1.668X5+0.109X6+1.053X7-0.439X8+1.068X9+1.796X10+1.573X11-0.802X12-0.258X13 (R2=0.918)，The linear regression equation of 0-6 month-old cows: Y=-182.399+0.97X1-0.359X2+0.579X3-0.097X4-0.683X5+0.025X6-0.364X7+0.502X8+0.329X9+1.374X10+1.786X11-0.083X12+0.097X13(R2=0.942)；The linear regression models based on four body measurements have the highest fitting degrees of 0.882 and 0.933, respectively. The linear regression equation of 0-6 month-old bulls is Y=-162.511+0.555X2+1.218X9+0.986X11-0.127X12(R2=0.882), and the linear regression equation of 0-6 month-old cows is Y=-175.941+0.436X2+0.449X9+2.073X11-0.213X12 (R2=0.933). Cattle of different sexes have different growth and development rates at different ages, except for bulls aged 13-18 months, which develop rapidly, change greatly and have obvious dynamics. The fitting effect of regression model between body size and weight is not very good, but it shows good fitting effect in other models of different ages and genders. 【 Conclusion 】 There is a significant correlation between body size and body weight of Xinjiang brown cattle, and a linear regression formula for predicting body weight based on body size is established. The model of cattle aged 0-6 months, 7-12 months and over 30 months shows that the predicted body weight and the measured body weight are not significant through paired sample T-test, which can be used to predict the body weight of cattle. This has a certain production practice for the growth and development detection and weight prediction of Xinjiang brown cattle at different ages and sexes.