Institute of Animal Science, University of Bonn, 53115 Bonn, Germany.
Faculty of Engineering, Albert Ludwig University of Freiburg, D-79110 Freiburg i. Br., Germany.
J Dairy Sci. 2022 Nov;105(12):9882-9895. doi: 10.3168/jds.2021-21547. Epub 2022 Oct 26.
The objective of the current study was to develop a predictive model for calf disease detection in the preweaning period using data from automated milk feeders (AMF). A deep convolutional neural network (CNN) architecture for the detection of respiratory disease and diarrhea in dairy calves was developed. German Holstein calves were fed milk replacer either ad libitum (up to 25 L/d; n = 32) or restrictively (6 L/d; n = 32) via AMF from 10 ± 3 d of life on. Concentrate, hay, and water were freely available. Calf health parameters were scored daily. The AMF measured milk replacer (MR) intake, number of rewarded visits, number of unrewarded visits, and drinking speed. A calf was considered sick if its fecal score was 3 or 4 and its respiratory score was 2 or 3. Only data from AMF up to 47 d of age were included in the analysis. This cut in the data was made to avoid data from the weaning period. Data were split in 80:20 ratios for training and testing data sets according to the Pareto principle. A minimum sensitivity of 80% was considered an appropriate requirement for the prediction models. Considering all calves in group housing, cross-validation of the test data set showed a sensitivity of 83% and a specificity of 79%, with a positive predictive value and a negative predictive value of 37 and 97%, respectively. The area under the curve of the receiver operating characteristic for the deep CNN model was 0.81 for all group-housed calves. The CNN model yielded sensitivity and specificity of 83 and 71%, respectively (for ad libitum-fed calves), and 82 and 87%, respectively (for restricted-fed calves), with good area under the curve-receiver operating characteristic (0.77 to 0.87), indicating that the CNN models can predict calf disease in both groups with different MR allowances. The permutation feature importance was measured by the decrease in model accuracy, and features (behaviors) were summarized in descending order of their relative importance to the CNN model. Drinking speed and MR intake were the main factors to predict calf disease in calves fed ad libitum. The number of unrewarded visits to the milk feeder and MR intake were the main factors to predict calf disease in restricted-fed calves. Despite the relatively small sample size, the results provide strong evidence that daily feeding behavior data from AMF can be used to identify calves at risk for disease. In conclusion, despite a very good testing performance of the CNN model, the relatively low daily prevalence of calf disease in the present study resulted in a high proportion of false-positive alarms.
本研究的目的是利用自动挤奶设备（AMF）的数据开发一种用于检测新生小牛疾病的预测模型。开发了一种用于检测奶牛犊呼吸疾病和腹泻的深度卷积神经网络（CNN）架构。从 10 ± 3 日龄开始，德国荷斯坦小牛通过 AMF 自由采食（高达 25 L/d；n = 32）或限制采食（6 L/d；n = 32）。浓缩物、干草和水均可自由获取。每日对小牛的健康参数进行评分。AMF 测量代乳料（MR）摄入量、奖励访问次数、未奖励访问次数和饮水速度。如果小牛的粪便评分为 3 或 4，呼吸评分为 2 或 3，则认为其患病。仅将 AMF 至 47 日龄的数据纳入分析。为避免断奶期的数据，在数据中进行了此切割。根据帕累托原则，将数据按 80:20 的比例分割为训练和测试数据集。预测模型的最小灵敏度为 80%被认为是适当的要求。考虑到所有分组饲养的小牛，对测试数据集进行的交叉验证显示，敏感性为 83%，特异性为 79%，阳性预测值和阴性预测值分别为 37%和 97%。深度 CNN 模型的接收者操作特征曲线下面积（AUC-ROC）为所有分组饲养小牛的 0.81。对于自由采食的小牛，CNN 模型的敏感性和特异性分别为 83%和 71%（对于限制采食的小牛），分别为 82%和 87%（对于限制采食的小牛），AUC-ROC 曲线的面积（0.77 至 0.87）表明，CNN 模型可以在允许不同 MR 的两组中预测小牛疾病。通过模型准确性的降低来衡量排列特征重要性，并按相对重要性降序对特征（行为）进行总结。对于自由采食的小牛，饮水速度和 MR 摄入量是预测小牛疾病的主要因素。对于限制采食的小牛，未奖励访问挤奶设备的次数和 MR 摄入量是预测小牛疾病的主要因素。尽管样本量相对较小，但结果提供了强有力的证据，表明来自 AMF 的日常喂养行为数据可用于识别患病风险较高的小牛。总之，尽管 CNN 模型的测试性能非常好，但本研究中小牛疾病的每日患病率相对较低，导致误报率较高。
