摘要
为了研究"精牧"系统检测奶牛发情效果,利用"精牧"奶牛独立发情检测系统检测试验所用的318头经产荷斯坦奶牛的发情数据,提取并保存,同时选派资深的发情检测人员对试验所用的318头牛只做直肠鉴定检测,并使用R语言(开源统计分析软件){gmodels}包中的CrossTable()函数和{caret}包中的confusionMatrix()函数,计算出"精牧"系统奶牛疑似发情检出率与正确率,利用混淆矩阵分析检验验证"精牧"奶牛独立发情检测系统检测数据与直肠鉴定检测数据的一致性。试验结果表明,"精牧"奶牛独立发情检测系统检测奶牛疑似发情检出率为0.957,正确率为0.903。Accuracy准确率为0.9465,Sensitivity(覆盖率或敏感度)值为0.9403,Specificity(负例覆盖率或特异度)值为0.9573,Pos Pred Value(命中率)值为0.9742和Neg Pred Value(负例命中率)值为0.9032,并且Kappa值为0.8865,充分说明"精牧"奶牛独立发情检测系统结果与直肠鉴定发情结果具有一致性。试验结果表明,"精牧"系统可以替代传统的鉴别方法鉴别奶牛是否发情。
In order to study the effect of "JING MU" system on detecting the estrus of dairy cows,the data of 318 multiparous Holstein cows in estrus were extracted and saved by using "JING MU'system,as well as the rectum of 318 Holstein cows were detection and identification by senior testers,repectively. The detection rate and accuracy rate of "JING MU"system werecalculated by CrossTable () function from {gmodels} packages and confusionMatrix() function from { caret} packages in R language. Moreover, the confusion matrix analysis test was used to verify the consistency of test data by using"JING MU" cow in- dependent estrus detection system and rectum identification. The results showed that the detection rate of suspected estrus was 0. 957 and the correct rate was 0. 903. Accuracy was 0. 9465, sensitivity was 0. 9403, specificity was 0. 9573,Positive Predictive Value was 0. 9742, Negative Predicted Value was 0. 9032, and the Kappa value was 0. 8865,which indicated that the results of the independent estrus detection system were consistent with the results of rectal identification. The results indicated that the "JING MU" system can replace the traditional identification method to identify whether the cow is estrus.
作者
徐斐飞
夏婷
徐刚
李春燕
刘继峰
XU Fei-fei XIA Ting XU Gang LI Chun-yan LIU Ji-feng(The Shandong Internet and Technology co. , L TD, Taian, Shandong, 271000)
出处
《中国牛业科学》
2017年第3期4-7,共4页
China Cattle Science
关键词
检出率
正确率
奶牛发情检测系统
R语言
混淆矩阵
KAPPA值
detection rate
correct rate
cow independent estrus detection system
R language
confusion matrix
Kappa value