A decision-tree model to detect post-calving diseases based on rumination, activity, milk yield, BW and voluntary visits to the milking robot

The Prediction of Clinical Mastitis in Dairy Cows Based on Milk Yield, Rumination Time, and Milk Electrical Conductivity Using Machine Learning Algorithms

In commercial dairy farms, mastitis is associated with increased antimicrobial use and associated resistance, which may affect milk production. This study aimed to develop sensor-based prediction models for naturally occurring clinical bovine mastitis using nine machine learning algorithms with data from 447 mastitic and 2146 healthy cows obtained from five commercial farms in Northeast China. The variables were related to daily activity, rumination time, and daily milk yield of cows, as well as milk electrical conductivity. Both Z-standardized and non-standardized datasets pertaining to four specific stages of lactation were used to train and test prediction models. For all four subgroups, the Z-standardized dataset yielded better results than those of the non-standardized one, with the multilayer artificial neural net algorithm showing the best performance. Variables of importance had a similar rank in this algorithm, indicating the consistency of these variables as predictors for bovine mastitis in commercial farms with similar automatic systems. Moreover, the peak milk yield (PMY) of mastitic cows was significantly higher than that of healthy cows (p < 0.005), indicating that high-yielding cattle are more prone to mastitis. Our results show that machine learning algorithms are effective tools for predicting mastitis in dairy cows for immediate intervention and management in commercial farms.

A Literature Review of Modeling Approaches Applied to Data Collected in Automatic Milking Systems

Automatic milking systems (AMS) have played a pioneering role in the advancement of Precision Livestock Farming, revolutionizing the dairy farming industry on a global scale. This review specifically targets papers that focus on the use of modeling approaches within the context of AMS. We conducted a thorough review of 60 articles that specifically address the topics of cows' health, production, and behavior/management Machine Learning (ML) emerged as the most widely used method, being present in 63% of the studies, followed by statistical analysis (14%), fuzzy algorithms (9%), deterministic models (7%), and detection algorithms (7%). A significant majority of the reviewed studies (82%) primarily focused on the detection of cows' health, with a specific emphasis on mastitis, while only 11% evaluated milk production. Accurate forecasting of dairy cow milk yield and understanding the deviation between expected and observed milk yields of individual cows can offer significant benefits in dairy cow management. Likewise, the study of cows' behavior and herd management in AMSs is under-explored (7%). Despite the growing utilization of machine learning (ML) techniques in the field of dairy cow management, there remains a lack of a robust methodology for their application. Specifically, we found a substantial disparity in adequately balancing the positive and negative classes within health prediction models.

Joint Models to Predict Dairy Cow Survival from Sensor Data Recorded during the First Lactation

Early predictions of cows' probability of survival to different lactations would help farmers in making successful management and breeding decisions. For this purpose, this research explored the adoption of joint models for longitudinal and survival data in the dairy field. An algorithm jointly modelled daily first-lactation sensor data (milk yield, body weight, rumination time) and survival data (i.e., time to culling) from 6 Holstein dairy farms. The algorithm was set to predict survival to the beginning of the second and third lactations (i.e., second and third calving) from sensor observations of the first 60, 150, and 240 days in milk of cows' first lactation. Using 3-time-repeated 3-fold cross-validation, the performance was evaluated in terms of Area Under the Curve and expected error of prediction. Across the different scenarios and farms, the former varied between 45% and 76%, while the latter was between 3.5% and 26%. Significant results were obtained in terms of expected error of prediction, meaning that the method provided survival probabilities in line with the observed events in the datasets (i.e., culling). Furthermore, the performances were stable among farms. These features may justify further research on the use of joint models to predict the survival of dairy cattle.

Industry 4.0 and Precision Livestock Farming (PLF): An up to Date Overview across Animal Productions

Precision livestock farming (PLF) has spread to various countries worldwide since its inception in 2003, though it has yet to be widely adopted. Additionally, the advent of Industry 4.0 and the Internet of Things (IoT) have enabled a continued advancement and development of PLF. This modern technological approach to animal farming and production encompasses ethical, economic and logistical aspects. The aim of this review is to provide an overview of PLF and Industry 4.0, to identify current applications of this rather novel approach in different farming systems for food producing animals, and to present up to date knowledge on the subject. Current scientific literature regarding the spread and application of PLF and IoT shows how efficient farm animal management systems are destined to become. Everyday farming practices (feeding and production performance) coupled with continuous and real-time monitoring of animal parameters can have significant impacts on welfare and health assessment, which are current themes of public interest. In the context of feeding a rising global population, the agri-food industry and industry 4.0 technologies may represent key features for successful and sustainable development.

The Early Prediction of Common Disorders in Dairy Cows Monitored by Automatic Systems with Machine Learning Algorithms

We use multidimensional data from automated monitoring systems and milking systems to predict disorders of dairy cows by employing eight machine learning algorithms. The data included the season, days in milking, parity, age at the time of disorders, milk yield (kg/day), activity (unitless), six variables related to rumination time, and two variables related to the electrical conductivity of milk. We analyze 131 sick cows and 149 healthy cows with identical lactation days and parity; all data are collected on the same day, which corresponds to the diagnosis day for disordered cows. For disordered cows, each variable, except the ratio of rumination time from daytime to nighttime, displays a decreasing/increasing trend from d-7 or d-3 to d0 and/or d-1, with the d0, d-1, or d-2 values reaching the minimum or maximum. The test data sensitivity for three algorithms exceeded 80%, and the accuracies of the eight algorithms ranged from 65.08% to 84.21%. The area under the curve (AUC) of the three algorithms was >80%. Overall, Rpart best predicts the disorders with an accuracy, precision, and AUC of 81.58%, 92.86%, and 0.908, respectively. The machine learning algorithms may be an appropriate and powerful decision support and monitoring tool to detect herds with common health disorders.

Assessment of feeding, ruminating and locomotion behaviors in dairy cows around calving – a retrospective clinical study to early detect spontaneous disease appearance

The study aims to verify the usefulness of new intervals-based algorithms for clinical interpretation of animal behavior in dairy cows around calving period. Thirteen activities associated with feeding-ruminating-locomotion-behaviors of 42 adult Holstein-Friesian cows were continuously monitored for the week (wk) -2, wk -1 and wk +1 relative to calving (overall 30’340 min/animal). Soon after, animals were retrospectively assigned to group-S (at least one spontaneous diseases; n = 24) and group-H (healthy; n = 18). The average activities performed by the groups, recorded by RumiWatch® halter and pedometer, were compared at the different weekly intervals. The average activities on the day of clinical diagnosis (dd0), as well as one (dd-1) and two days before (dd-2) were also assessed. Differences of dd0 vs. dd-1 (ΔD1), dd0 vs. wk -1 (ΔD2), and wk +1 vs. wk -1 (Δweeks) were calculated. Variables showing significant differences between the groups were used for a univariate logistic regression, a receiver operating characteristic analysis, and a multivariate logistic regression model. At wk +1 and dd0, eating- and ruminating-time, eating- and ruminate-chews and ruminating boluses were significantly lower in group-S as compared to group-H, while other activity time was higher. For ΔD2 and Δweeks, the differences of eating- and ruminating-time, as well as of eating-and ruminate-chews were significantly lower in group-S as compared to group-H. Concerning the locomotion behaviors, the lying time was significantly higher in group-S vs. group-H at wk +1 and dd-2. The number of strides was significantly lower in group-S compared to group-H at wk +1. The model including eating-chews, ruminate-chews and other activity time reached the highest accuracy in detecting sick cows in wk +1 (area under the curve: 81%; sensitivity: 73.7%; specificity: 82.4%). Some of the new algorithms for the clinical interpretation of cow behaviour as described in this study may contribute to monitoring animals’ health around calving.

Comparison of various classification techniques for supervision of milk processing

Detecting the types of anomalies that can occur throughout the milk processing process is an important task since it can assist providers in maintaining control over the process. The Raman spectrometer was used in conjunction with several classification approaches—linear discriminant analysis, decision tree, support vector machine, and k nearest neighbor—to establish a viable method for detecting different types of anomalies that may occur during the process—temperature and fat variation and added water or cleaning solution. Milk with 5% fat measured at 10°C was used as the reference milk for this study. Added water, cleaning solution, milk with various fat contents and different temperatures were used to detect abnormal conditions. While decision trees and linear discriminant analysis were unable to accurately categorize the various type of anomalies, the k nearest neighbor and support vector machine provided promising results. The accuracy of the support vector machine test set and the k nearest neighbor test set were 81.4% and 84.8%, respectively. As a result, it is reasonable to conclude that both algorithms are capable of appropriately classifying the various groups of samples. It can assist milk industries in determining what is wrong during milk processing.

Exploring the Potential of Precision Livestock Farming Technologies to Help Address Farm Animal Welfare

The rise in the demand for animal products due to demographic and dietary changes has exacerbated difficulties in addressing societal concerns related to the environment, human health, and animal welfare. As a response to this challenge, Precision Livestock Farming (PLF) technologies are being developed to monitor animal health and welfare parameters in a continuous and automated way, offering the opportunity to improve productivity and detect health issues at an early stage. However, ethical concerns have been raised regarding their potential to facilitate the management of production systems that are potentially harmful to animal welfare, or to impact the human-animal relationship and farmers' duty of care. Using the Five Domains Model (FDM) as a framework, the aim is to explore the potential of PLF to help address animal welfare and to discuss potential welfare benefits and risks of using such technology. A variety of technologies are identified and classified according to their type [sensors, bolus, image or sound based, Radio Frequency Identification (RFID)], their development stage, the species they apply to, and their potential impact on welfare. While PLF technologies have promising potential to reduce the occurrence of diseases and injuries in livestock farming systems, their current ability to help promote positive welfare states remains limited, as technologies with such potential generally remain at earlier development stages. This is likely due to the lack of evidence related to the validity of positive welfare indicators as well as challenges in technology adoption and development. Finally, the extent to which welfare can be improved will also strongly depend on whether management practices will be adapted to minimize negative consequences and maximize benefits to welfare.

Do we automatically detect health- or general welfare-related issues? A framework

The early detection of health disorders is a central goal in livestock production. Thus, a great demand for technologies enabling the automated detection of such issues exists. However, despite decades of research, precision livestock farming (PLF) technologies with sufficient accuracy and ready for implementation on commercial farms are rare. A central factor impeding technological development is likely the use of non-specific indicators for various issues. On commercial farms, where animals are exposed to changing environmental conditions, where they undergo different internal states and, most importantly, where they can be challenged by more than one issue at a time, such an approach leads inevitably to errors. To improve the accuracy of PLF technologies, the presented framework proposes a categorization of the aim of detection of issues related to general welfare, disease and distress and defined disease. Each decision level provides a different degree of information and therefore requires indicators varying in specificity. Based on these considerations, it becomes apparent that while most technologies aim to detect a defined health issue, they facilitate only the identification of issues related to general welfare. To achieve detection of specific issues, new indicators such as rhythmicity patterns of behaviour or physiological processes should be examined.

Variable selection for monitoring sickness behavior in lactating dairy cattle with the application of control charts

The present observational study investigated the application of multivariate cumulative sum (MCUSUM) control charts by including variables selected by principal component analysis and partial least squares (PLS) regression to identify sickness behavior in dairy cattle. Therefore, sensor information (24 variables) was collected from 480 milking cows on a German dairy farm between September 2018 and December 2019. These variables were gathered in potentially different scenarios on farm. In total, data from 749 animals were available for evaluation. Variables were chosen based on the information of 499 cows (62 healthy; 437 sick) with 93,598 observations. The available diagnoses were collected together to form 1,025 sickness events. Hence, the different numbers of selected variables were included into the MCUSUM control charts. The performance of the MCUSUM control charts was evaluated by a 10-fold cross validation; hence, 90% of the original data set (749 cows) represented the training data, and the remaining 10% was used to test the training results. On average, the 10 training data sets included 124,871 observations with 1,392 sickness events, and the 10 testing data sets included, on average, 13,704 observations with 153 sickness events. The MCUSUM generated from the variables selected by principal component analysis showed comparable results in training and testing in all scenarios; therefore, 70.0 to 97.4% of the sickness events were detected. The false-positive rates ranged from 8.5 to 29.6%, and thus they created at least 2.6 false-positive alerts per day in testing. The variables selected by the PLS regression approach showed comparable sickness detection rates (70.0-99.9%) as well as false-positive rates (8.2-62.8%) in most scenarios. The best performing scenario produced 2.5 false-positive alerts in testing. Summarizing, both approaches showed potential for practical implementation; however, the PLS variable selection approach showed fewer false positives. Therefore, the PLS regression approach could generate a more reliable sickness detection algorithm, if combined with MCUSUM control charts, and considered for practical implementation.

Advancements in sensor technology and decision support intelligent tools to assist smart livestock farming

Remote monitoring, modern data collection through sensors, rapid data transfer, and vast data storage through the Internet of Things (IoT) have advanced precision livestock farming (PLF) in the last 20 yr. PLF is relevant to many fields of livestock production, including aerial- and satellite-based measurement of pasture’s forage quantity and quality; body weight and composition and physiological assessments; on-animal devices to monitor location, activity, and behaviors in grazing and foraging environments; early detection of lameness and other diseases; milk yield and composition; reproductive measurements and calving diseases; and feed intake and greenhouse gas emissions, to name just a few. There are many possibilities to improve animal production through PLF, but the combination of PLF and computer modeling is necessary to facilitate on-farm applicability. Concept- or knowledge-driven (mechanistic) models are established on scientific knowledge, and they are based on the conceptualization of hypotheses about variable interrelationships. Artificial intelligence (AI), on the other hand, is a data-driven approach that can manipulate and represent the big data accumulated by sensors and IoT. Still, it cannot explicitly explain the underlying assumptions of the intrinsic relationships in the data core because it lacks the wisdom that confers understanding and principles. The lack of wisdom in AI is because everything revolves around numbers. The associations among the numbers are obtained through the “automatized” learning process of mathematical correlations and covariances, not through “human causation” and abstract conceptualization of physiological or production principles. AI starts with comparative analogies to establish concepts and provides memory for future comparisons. Then, the learning process evolves from seeking wisdom through the systematic use of reasoning. AI is a relatively novel concept in many science fields. It may well be “the missing link” to expedite the transition of the traditional maximizing output mentality to a more mindful purpose of optimizing production efficiency while alleviating resource allocation for production. The integration between concept- and data-driven modeling through parallel hybridization of mechanistic and AI models will yield a hybrid intelligent mechanistic model that, along with data collection through PLF, is paramount to transcend the current status of livestock production in achieving sustainability.

Smart and Climate-Smart Agricultural Trends as Core Aspects of Smart Village Functions

Attention has shifted to the development of villages in Europe and other parts of the world with the goal of combating rural–urban migration, and moving toward self-sufficiency in rural areas. This situation has birthed the smart village idea. Smart village initiatives such as those of the European Union is motivating global efforts aimed at improving the live and livelihood of rural dwellers. These initiatives are focused on improving agricultural productivity, among other things, since most of the food we eat are grown in rural areas around the world. Nevertheless, a major challenge faced by proponents of the smart village concept is how to provide a framework for the development of the term, so that this development is tailored towards sustainability. The current work examines the level of progress of climate smart agriculture, and tries to borrow from its ideals, to develop a framework for smart village development. Given the advances in technology, agricultural development that encompasses reduction of farming losses, optimization of agricultural processes for increased yield, as well as prevention, monitoring, and early detection of plant and animal diseases, has now embraced varieties of smart sensor technologies. The implication is that the studies and results generated around the concept of climate smart agriculture can be adopted in planning of villages, and transforming them into smart villages. Hence, we argue that for effective development of the smart village framework, smart agricultural techniques must be prioritized, viz-a-viz other developmental practicalities.

A survey on sensor systems used in Italian dairy farms and comparison between performances of similar herds equipped or not equipped with sensors

Sensor systems (SS) were developed over the last few decades to help dairy farmers manage their herds. Such systems can provide both data and alerts to several productive, behavioral, and physiological indicators on individual cows. Currently, there is still a lack of knowledge on both the proportion of dairy farms that invested in SS and type of SS installed. Additionally, it is still unclear whether the performances of herds equipped with SS differ from those of similar herds managed without any technological aid. Therefore, the aims of this study were (1) to provide an insight into SS spread among Italian dairy farms and (2) to analyze the performances of similar herds equipped or not equipped with SS. To reach the former goal, a large survey was carried out on 964 dairy farms in the northeast of Italy. Farmers were interviewed by the technicians of the regional breeders association to collect information on the type of SS installed on farms and the main parameters recorded. Overall, 42% of the surveyed farms had at least 1 SS, and most of them (72%) reared more than 50 cows. Sensors for measuring individual cow milk yield were the most prevalent type installed (39% of the surveyed farms), whereas only 15% of farms had SS for estrus detection. More sophisticated parameters, such as rumination, were automatically monitored in less than 5% of the farms. To reach the latter goal of the study, a subset of 100 Holstein dairy farms with similar characteristics was selected: half of them were equipped with SS for monitoring at least individual milk yield and estrus, and the other half were managed without any SS. Average herd productive and reproductive data from official test days over 3 yr were analyzed. The outcomes of the comparison showed that farms with SS had higher mature-equivalent milk production. Further clustering analysis of the same 100 farms partitioned them into 3 clusters based on herd productive and reproductive data. Results of the Chi-squared test showed that the proportion of farms equipped with SS was greater in the cluster with the best performance (e.g., higher milk yield and shorter calving interval). However, the presence of a few farms equipped with SS in the least productive cluster for the same parameters pointed out that although the installation of SS may support farmers in time- and labor-saving or in data recording, it is not a guarantee of better herd performance.

Potential role of MicroRNA as a diagnostic tool in the detection of bovine mastitis

Bovine mastitis is a major health problem that affects dairy cows and has a negative impact on milk production. The presence of microRNAs in biofluids, such as blood and milk, could play a pivotal role in the detection of bovine mastitis. The purpose of the current study was to determine the levels of microRNA gene expression in milk, in combination with other reported mastitis indicators, as a biomarker of bovine mastitis. Milk samples (n = 171) were obtained from 113 dairy cows with known disease status (i.e., healthy; n = 23 cows, subclinical mastitis; n = 45 cows, or clinical mastitis; n = 45 cows) and analyzed for the presence of MIR24-2, MIR29B-2, MIR146A, MIR148A, MIR155, MIR181A1, MIR184, and MIR223 expression using the real-time PCR (qPCR) method. The expression data were then utilized in the creation of receiver operator characteristic curves (ROC) and further analyzed by the machine learning (ML) methods. MIR29B-2, MIR146A, MIR148A, and MIR155 expression levels differed significantly among the three groups. These potential microRNA biomarkers of mastitis exhibited high sensitivity and specificity. Next, we applied ML algorithm, specifically, a decision tree (DT) model to predict the status of milk based on MIR29B-2 and MIR146A expression levels. The results suggested that MIR29B-2, when used in combination with the California mastitis test (CMT) and days in milk (DIM) data, was applicable for screening and classification of milk samples from cows as healthy, subclinical mastitis, or mastitis. MIR29B-2 appears to have sufficient discriminatory power to enable it to be utilized as a biomarker in cases where the status of a milk sample cannot be determined based on CMT results.

Forecasting the milk yield of cows on farms equipped with automatic milking system with the use of decision trees

The purpose of this paper was to utilize the decision trees technique to determine the factors responsible for high monthly milk yield in Polish Holstein-Friesian cows from 27 herds equipped with milking robots. The applied statistical method-the decision tree technique-showed that the most important factors responsible for monthly milk yield of dairy cows using robots were, in descending order of importance: milking frequency, lactation number, month of milking, and type of lying stall. At the same time, it has been ascertained that the highest monthly milk yield (47.24 kg) can be expected from multiparous cows kept in barns with a deep bedding that were milked more frequently than three times per day. On the other hand, the lowest milk production (13.56 kg) was observed among dairy cows milked less frequently than two times a day, with an average number of milked quarters lower than 3.97. The application of the decision trees technique allows a breeder to select appropriate levels of environmental factors and parameters that will help to ensure maximized milk production.

Machine Learning Based Prediction of Insufficient Herbage Allowance with Automated Feeding Behaviour and Activity Data

Sensor technologies that measure grazing and ruminating behaviour as well as physical activities of individual cows are intended to be included in precision pasture management. One of the advantages of sensor data is they can be analysed to support farmers in many decision-making processes. This article thus considers the performance of a set of RumiWatchSystem recorded variables in the prediction of insufficient herbage allowance for spring calving dairy cows. Several commonly used models in machine learning (ML) were applied to the binary classification problem, i.e., sufficient or insufficient herbage allowance, and the predictive performance was compared based on the classification evaluation metrics. Most of the ML models and generalised linear model (GLM) performed similarly in leave-out-one-animal (LOOA) approach to validation studies. However, cross validation (CV) studies, where a portion of features in the test and training data resulted from the same cows, revealed that support vector machine (SVM), random forest (RF) and extreme gradient boosting (XGBoost) performed relatively better than other candidate models. In general, these ML models attained 88% AUC (area under receiver operating characteristic curve) and around 80% sensitivity, specificity, accuracy, precision and F-score. This study further identified that number of rumination chews per day and grazing bites per minute were the most important predictors and examined the marginal effects of the variables on model prediction towards a decision support system.

Prospects for predictive modeling of transition cow diseases

Transition cow diseases can negatively impact animal welfare and reduce dairy herd profitability. Transition cow disease incidence has remained relatively stable over time despite monitoring and management efforts aimed to reduce the risk of developing diseases. Dairy cattle disease risk is monitored by assessing multiple factors, including certain biomarker test results, health records, feed intake, body condition score, and milk production. However, these factors, which are used to make herd management decisions, are often reviewed separately without considering the correlation between them. In addition, the biomarkers that are currently used for monitoring may not be representative of the complex physiological changes that occur during the transition period. Predictive modeling, which uses data to predict future or current outcomes, is a method that can be used to combine the most predictive variables and their interactions efficiently. The use of an effective predictive model with relevant predictors for transition cow diseases will result in better targeted interventions, and therefore lower disease incidence. This review will discuss predictive modeling methods and candidate variables in the context of transition cow diseases. The next step is to investigate novel biomarkers and statistical methods that are best suited for the prediction of transition cow diseases.

Smart Animal Agriculture: Application of Real-Time Sensors to Improve Animal Well-Being and Production

Consumption of animal products such as meat, milk, and eggs in first-world countries has leveled off, but it is rising precipitously in developing countries. Agriculture will have to increase its output to meet demand, opening the door to increased automation and technological innovation; intensified, sustainable farming; and precision livestock farming (PLF) applications. Early indicators of medical problems, which use sensors to alert cattle farmers early concerning individual animals that need special care, are proliferating. Wearable technologies dominate the market. In less-value-per-animal systems like sheep, goat, pig, poultry, and fish, one sensor, like a camera or robot per herd/flock/school, rather than one sensor per animal, will become common. PLF sensors generate huge amounts of data, and many actors benefit from PLF data. No standards currently exist for sharing sensor-generated data, limiting the use of commercial sensors. Technologies providing accurate data can enhance a well-managed farm. Development of methods to turn the data into actionable solutions is critical.

Rumination time and monitoring of health disorders during early lactation

The objective was to evaluate the association between changes in daily rumination time (dRT) and early stages of disease during early lactation and to assess the performance of two proposed disease detection indices. This cohort study included 210 multiparous Holstein cows at the University of Florida Dairy Unit. Cows were affixed with a neck collar containing rumination loggers providing rumination time. The occurrence of health disorders (mastitis, metritis, clinical hypocalcemia, depression/dehydration/fever (DDF), digestive conditions, lameness and clinical ketosis) was assessed until 60 days in milk. Two indices were developed to explore the association between dRT and disease: (i) Cow index (CIx), based on changes in dRT in the affected cow during the days before the diagnosis of health disorders; (ii) Mates index (MIx), index based on deviations in dRT relative to previous days and healthy pen mate cohorts. Subsequently, an alert model was proposed for each index (ACIx and AMIx) considering the reference alert cut-off values as the differences between average index values in healthy and sick cows for each specific disease. The sensitivity (SE) of ACIx detecting disease ranged from 42% (digestive conditions and DDF) to 80% (clinical hypocalcemia) with 84% specificity (SP). The SE of AMIx ranged from 46% (digestive conditions and DDF) to 100% (clinical hypocalcemia) with 85% SP. Consistent reductions in rumination activity, both within cow and relative to healthy mate cohorts, were observed for each health disorder at the day of diagnosis. However, the ability of the proposed algorithms for detecting each specific disease was variable.

Towards practical application of sensors for monitoring animal health; design and validation of a model to detect ketosis

The objective of this study was to design and validate a mathematical model to detect post-calving ketosis. The validation was conducted in four commercial dairy farms in Israel, on a total of 706 multiparous Holstein dairy cows: 203 cows clinically diagnosed with ketosis and 503 healthy cows. A logistic binary regression model was developed, where the dependent variable is categorical (healthy/diseased) and a set of explanatory variables were measured with existing commercial sensors: rumination duration, activity and milk yield of each individual cow. In a first validation step (within-farm), the model was calibrated on the database of each farm separately. Two thirds of the sick cows and an equal number of healthy cows were randomly selected for model validation. The remaining one third of the cows, which did not participate in the model validation, were used for model calibration. In order to overcome the random selection effect, this procedure was repeated 100 times. In a second (between-farms) validation step, the model was calibrated on one farm and validated on another farm. Within-farm accuracy, ranging from 74 to 79%, was higher than between-farm accuracy, ranging from 49 to 72%, in all farms. The within-farm sensitivities ranged from 78 to 90%, and specificities ranged from 71 to 74%. The between-farms sensitivities ranged from 65 to 95%. The developed model can be improved in future research, by employing other variables that can be added; or by exploring other models to achieve greater sensitivity and specificity.

Towards practical application of sensors for monitoring animal health: the effect of post-calving health problems on rumination duration, activity and milk yield

Three sources of sensory data: cow's individual rumination duration, activity and milk yield were evaluated as possible indicators for clinical diagnosis, focusing on post-calving health problems such as ketosis and metritis. Data were collected from a computerised dairy-management system on a commercial dairy farm with Israeli Holstein cows. In the analysis, 300 healthy and 403 sick multiparous cows were studied during the first 3 weeks after calving. A mixed model with repeated measurements was used to compare healthy cows with sick cows. In the period from 5 d before diagnosis and treatment to 2 d after it, rumination duration and activity were lower in the sick cows compared to healthy cows. The milk yield of sick cows was lower than that of the healthy cows during a period lasting from 5 d before until 5 d after the day of diagnosis and treatment. Differences in the milk yield of sick cows compared with healthy cows became greater from 5 to 1 d before diagnosis and treatment. The greatest significant differences occurred 3 d before diagnosis for rumination duration and 1 d before diagnosis for activity and milk yield. These results indicate that a model can be developed to automatically detect post-calving health problems including ketosis and metritis, based on rumination duration, activity and milk yield.