Comprehensive analysis of machine learning models for prediction of sub-clinical mastitis: Deep Learning and Gradient-Boosted Trees outperform other models

Ebrahimi, Mansour; Mohammadi‐Dehcheshmeh, Manijeh; Ebrahimie, Esmaeil; Petrovski, Kiro

doi:10.1016/j.compbiomed.2019.103456

Cited by 102 publications

(52 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…With the advent of increased "big data" within farm animal medicine, the potential to translate this into "smart data" is increasing 26 ; making full use of data already being collected. Machine learning has been applied to epidemiological classification problems within cattle medicine, such as the prediction of bovine viral diarrhoea virus exposure at herd level 27 , and the distribution of exposure of herds to liver fluke 28 , and has recently been applied in the investigation of mastitis pathogen (Streptococcus uberis) transmission patterns in cattle 29 as well as in the diagnosis of both subclinical 30,31 and clinical 32 mastitis at an individual animal level. Whilst machine learning has been described in both medical and veterinary fields and has been explored in the individual diagnosis of mastitis, it has not yet been applied to accurately replicate a specialist clinical diagnosis for a population level diagnosis, in this case the herd level diagnosis of bovine mastitis.…”

mentioning

confidence: 99%

Automated prediction of mastitis infection patterns in dairy herds using machine learning

Hyde

Down

Bradley

et al. 2020

Sci Rep

View full text Add to dashboard Cite

Mastitis in dairy cattle is extremely costly both in economic and welfare terms and is one of the most significant drivers of antimicrobial usage in dairy cattle. A critical step in the prevention of mastitis is the diagnosis of the predominant route of transmission of pathogens into either contagious (cont) or environmental (ENV), with environmental being further subdivided as transmission during either the nonlactating "dry" period (EDP) or lactating period (EL). Using data from 1000 farms, random forest algorithms were able to replicate the complex herd level diagnoses made by specialist veterinary clinicians with a high degree of accuracy. An accuracy of 98%, positive predictive value (PPV) of 86% and negative predictive value (NPV) of 99% was achieved for the diagnosis of CONT vs ENV (with CONT as a "positive" diagnosis), and an accuracy of 78%, PPV of 76% and NPV of 81% for the diagnosis of EDP vs EL (with EDP as a "positive" diagnosis). An accurate, automated mastitis diagnosis tool has great potential to aid non-specialist veterinary clinicians to make a rapid herd level diagnosis and promptly implement appropriate control measures for an extremely damaging disease in terms of animal health, productivity, welfare and antimicrobial use.

show abstract

mentioning

confidence: 99%

Automated prediction of mastitis infection patterns in dairy herds using machine learning

Hyde

Down

Bradley

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Most commonly, the datasets contain highly diverse features in sizes, units, and range. Since most of the machine learning models use Euclidean distance, it can affect the performance of the models [43,44]. The range of ICU dataset points used in this paper is widely varied; therefore, feature scaling is necessary to suppress the mentioned effects on the performance of models.…”

Section: ) Feature Scalingmentioning

confidence: 99%

A New Hybrid Predictive Model to Predict the Early Mortality Risk in Intensive Care Units on a Highly Imbalanced Dataset

et al. 2020

View full text Add to dashboard Cite

Due to the development of biomedical equipment and healthcare level, especially in the Intensive Care Unit (ICU), a considerable amount of data has been collected for analysis. Mortality prediction in the ICUs is considered as one of the most important topics in the healthcare data analysis section. A precise prediction of the mortality risk for patients in ICU could provide us with valuable information about patients' lives and reduce costs at the earliest possible stage. This paper aims to introduce a new hybrid predictive model using the Genetic Algorithm as a feature selection method and a new ensemble classifier based on the combination of Stacking and Boosting ensemble methods to create an early mortality prediction model on a highly imbalanced dataset. The SVM-SMOTE method is used to solve the imbalanced data problem. This paper compares the new model with various machine learning models to validate the efficiency of the introduced model. The achieved results using the shuffle 5-fold cross-validation and random hold-out methods indicate that the new hybrid model has the best performance among other classifiers. Additionally, the Friedman test is applied as a statistical significance test to examine the differences between classifiers. The results of the statistical analysis prove that the proposed model is more effective than other classifiers. Furthermore, the proposed model is compared to APACHE and SAPS scoring systems and is benchmarked against state-of-the-art predictive models applied to the MIMIC dataset for experimental validation and achieved promising results as it outperformed the state-of-the-art models.

show abstract

“…Machine learning and big data analytics have the potential to improve welfare and productivity in dairy cattle. They can be used to monitor and predict lameness and mastitis in dairy cattle, huge welfare issues that can have severe negative consequences on milk production (Ebrahimi et al, 2019;Taneja et al, 2020;Warner et al, 2020).…”

Section: Big Datamentioning

confidence: 99%

Digitalization of Animal Farming

Neethirajan¹

2020

Preprint

View full text Add to dashboard Cite

As the global human population increases, animal agriculture must adapt to provide more animal products while also addressing concerns about animal welfare, environmental sustainability, and public health. The purpose of this review is to discuss the digitalization of animal farming with Precision Livestock Farming (PLF) technologies, specifically biosensors, big data, and block chain technology. Biosensors are noninvasive or invasive sensors that monitor an animal’s health and behavior in real time, allowing farmers to monitor individual animals and integrate this data for population-level analyses. The data from the sensors is processed using big data-processing techniques such as data modelling. These technologies use algorithms to sort through large, complex data sets to provide farmers with biologically relevant and usable data. Blockchain technology allows for traceability of animal products from farm to table, a key advantage in monitoring disease outbreaks and preventing related economic losses and food-related health pandemics. With these PLF technologies, animal agriculture can become more transparent and regain consumer trust. While the digitalization of animal farming has the potential to address a number of pressing concerns, these technologies are relatively new. The implementation of PLF technologies on farms will require increased collaboration between farmers, animal scientists, and engineers to ensure that technologies can be used in realistic, on-farm conditions. These technologies will call for data models that can sort through large amounts of data while accounting for specific variables and ensuring automation, accessibility, and accuracy of data. Issues with data privacy, security, and integration will need to be addressed before there can be multi-farm databases. Lastly, the usage of blockchain technology in animal agriculture is still in its infancy; blockchain technology has the potential to improve the traceability and transparency of animal products, but more research is needed to realize its full potential. The digitalization of animal farming can supply the necessary tools to provide sustainable animal products on a global scale.

show abstract

Comprehensive analysis of machine learning models for prediction of sub-clinical mastitis: Deep Learning and Gradient-Boosted Trees outperform other models

Cited by 102 publications

References 19 publications

Automated prediction of mastitis infection patterns in dairy herds using machine learning

Automated prediction of mastitis infection patterns in dairy herds using machine learning

A New Hybrid Predictive Model to Predict the Early Mortality Risk in Intensive Care Units on a Highly Imbalanced Dataset

Digitalization of Animal Farming

Contact Info

Product

Resources

About