Prediction Of Clinical Mastitis In Dairy Cows Using ANN

Abstract

The ability of veterinarians to diagnose dairy cow diseases quickly and effectively is critical in dairy herd management. Electronic medical records have been utilised extensively to support clinical decisions for people utilising deep learning (DL). But veterinary diagnostics rarely use this technique. Furthermore, large datasets fuel most deep learning algorithms, disregarding the subjective knowledge acquired by veterinarians that is essential for illness diagnosis. This research suggests a DL approach for diagnosing dairy cow disease in order to address these problems: Artificial Neural Network (ANN). The model is trained using a dataset with parameters like Cow ID, Front Left Udder Inhale Limit, Day, Months after having given birth, breed, Front Left Udder Exhale Limit (IUFL), past incident of mastitis, Front Right Udder Exhale Limit(EUFR), Rear Left Udder Exhale Limit, Rear Right Udder Inhale Limit, Rear Right Udder Exhale Limit, cow temperature, udder hardness from manual intervention via switch, Front Right Udder Inhale Limit (IUFR), and pain from udder bloating. Mastitis cows are designated as 1, whereas normal cows are designated as 0. Experiments on dairy cow clinical datasets were carried out to validate its performance. The result show that our model performed well in disease diagnosis, with an accuracy rating of 99.090%, precision score of  99.100%, recall score of  99.100% and F1- score of  99%.The findings of the study have a significant impact on the fruitful, quick, and automated medical detection of illnesses in dairy cows.