Bayesian integration of sensor information and a multivariate dynamic linear model for prediction of dairy cow mastitis

Abstract

Rapid detection of dairy cow mastitis is important so corrective action can be taken as soon as possible. Automatically collected sensor data used to monitor the performance and the health state of the cow could be useful for rapid detection of mastitis while reducing the labor needs for monitoring. The state of the art in combining sensor data to predict clinical mastitis still does not perform well enough to be applied in practice. Our objective was to combine a multivariate dynamic linear model (DLM) with a naïve Bayesian classifier (NBC) in a novel method using sensor and nonsensor data to detect clinical cases of mastitis. We also evaluated reductions in the number of sensors for detecting mastitis. With the DLM, we co-modeled 7 sources of sensor data (milk yield, fat, protein, lactose, conductivity, blood, body weight) collected at each milking for individual cows to produce one-step-ahead forecasts for each sensor. The observations were subsequently categorized according to the errors of the forecasted values and the estimated forecast variance. The categorized sensor data were combined with other data pertaining to the cow (week in milk, parity, mastitis history, somatic cell count category, and season) using Bayes' theorem, which produced a combined probability of the cow having clinical mastitis. If this probability was above a set threshold, the cow was classified as mastitis positive. To illustrate the performance of our method, we used sensor data from 1,003,207 milkings from the University of Florida Dairy Unit collected from 2008 to 2014. Of these, 2,907 milkings were associated with recorded cases of clinical mastitis. Using the DLM/NBC method, we reached an area under the receiver operating characteristic curve of 0.89, with a specificity of 0.81 when the sensitivity was set at 0.80. Specificities with omissions of sensor data ranged from 0.58 to 0.81. These results are comparable to other studies, but differences in data quality, definitions of clinical mastitis, and time windows make comparisons across studies difficult. We found the DLM/NBC method to be a flexible method for combining multiple sensor and nonsensor data sources to predict clinical mastitis and accommodate missing observations. Further research is needed before practical implementation is possible. In particular, the performance of our method needs to be improved in the first 2 wk of lactation. The DLM method produces forecasts that are based on continuously estimated multivariate normal distributions, which makes forecasts and forecast errors easy to interpret, and new sensors can easily be added.

Identification of Key Candidate Genes in Dairy Cow in Response to Escherichia coli Mastitis by Bioinformatical Analysis

“…Early diagnosis is very important due to the high costs of mastitis. However, most current detection systems do not meet the high accuracy required for clinical diagnosis needs of mastitis (Jensen et al, 2016). Nowadays, somatic cell count (SCC) and California mastitis test (CMT) are often used in the diagnosis of mastitis.…”

“…In view of the fact that the performances of the most current bovine mastitis detection systems do not meet the high accuracy required for clinical diagnosis needs of mastitis (Jensen et al, 2016), the main purpose of this study is to identify potential genes that can be used as biomarkers for the diagnosis of E. coli mastitis. Although these genes have been identified in this study and had been preliminarily confirmed in previous animal or cell Cytokine-cytokine receptor interaction 8.26E−09 CCL3, IL6, CCL2, CCL20, IL18, IL1B, CXCL8, IL1A bta04621 NOD-like receptor signaling pathway 6.70E−08 IL6, CCL2, IL18, IL1B, NFKBIA, CXCL8 bta04620…”

“…Therefore, it is difficult to use differential genes obtained in a single independent study as biomarkers of E. coli mastitis. So far, the performance of the most mastitis detection systems do not meet the high accuracy required for clinical diagnosis needs of mastitis in cattle (Jensen et al, 2016). The aim of developmental research is to enhance the diagnostic efficiency of bovine mastitis with including several biomarkers on one test strip.…”

Comparing regression, naive Bayes, and random forest methods in the prediction of individual survival to second lactation in Holstein cattle

“…The naive Bayes method relies on independence between the input variables, but it performs surprisingly well even under conditions that might be considered suboptimal for the algorithm (Domingos and Pazzani, 1997;Friedman, 1997). Despite the relative simplicity of its algorithm, naive Bayes is still widely used (Jensen et al, 2016;Drury et al, 2017). Random forest (Breiman, 2001) is another machine-learning method that is successfully implemented in a wide variety of fields, including animal science (Shahinfar et al, 2014;Machado et al, 2015;Brieuc et al, 2018).…”

Multivariate dynamic linear models for estimating the effect of experimental interventions in an evolutionary operations setup in dairy herds

“…Dynamic linear model have been used in monitoring bulk tank SCC in milk (Thysen, 1993), in detecting estrus and diseases in dairy cattle (de Mol et al, 1999), and in monitoring daily milk production (Van Bebber et al, 1999). In recent years, DLM have been used to estimate the production potential of cows as a basis for optimal replacement policies (Nielsen et al, 2010), for prediction of dairy cow mastitis (Jensen et al, 2016), as well as to determine whether dairy cows with subclinical mammary infections recover after antibiotic treatment (Jørgensen et al, 2016). The model constructed by Van Bebber et al (1999) was used to detect significant changes in the milk production of both individual cows and an entire herd, and could be used in exploiting changes occurring during normal (non-EVOP) production flow in the dairy herd.…”