Classifying Chewing and Rumination in Dairy Cows Using Sound Signals and Machine Learning

SIMPLE SUMMARY: This article aims to investigate the nutritional behavior of dairy cattle, aiming to comprehend their dietary requirements and eating habits. In this regard, an effort has been made to scrutinize dietary patterns by analyzing sound recordings captured from the cows’ jaws during the chewing process. The paper outlines the methodology for developing various models to discern nutritional patterns in dairy cattle, employing six well-known classifiers. Understanding nutritional behavior and dietary patterns in dairy cattle is crucial for livestock managers and animal welfare. By comprehending the dietary requirements and eating habits of dairy cattle, managers can ensure that the cows are receiving the appropriate nutrients to maintain their health and productivity. This information can also help managers identify any potential health issues or deficiencies in the cows’ diets, allowing for early intervention and prevention of further health problems. Additionally, understanding the nutritional behavior of dairy cattle can lead to more efficient feeding practices, reducing waste and costs associated with overfeeding or underfeeding. Ultimately, establishing an appropriate pattern for evaluating the nutrition of dairy cattle can serve as a valuable guide for livestock managers to ensure the well-being and welfare of the cows while also improving the overall productivity and profitability of the dairy farm. ABSTRACT: This research paper introduces a novel methodology for classifying jaw movements in dairy cattle into four distinct categories: bites, exclusive chews, chew-bite combinations, and exclusive sorting, under conditions of tall and short particle sizes in wheat straw and Alfalfa hay feeding. Sound signals were recorded and transformed into images using a short-time Fourier transform. A total of 31 texture features were extracted using the gray level co-occurrence matrix, spatial gray level dependence method, gray level run length method, and gray level difference method. Genetic Algorithm (GA) was applied to the data to select the most important features. Six distinct classifiers were employed to classify the jaw movements. The total precision found was 91.62%, 94.48%, 95.9%, 92.8%, 94.18%, and 89.62% for Naive Bayes, k-nearest neighbor, support vector machine, decision tree, multi-layer perceptron, and k-means clustering, respectively. The results of this study provide valuable insights into the nutritional behavior and dietary patterns of dairy cattle. The understanding of how cows consume different types of feed and the identification of any potential health issues or deficiencies in their diets are enhanced by the accurate classification of jaw movements. This information can be used to improve feeding practices, reduce waste, and ensure the well-being and productivity of the cows. The methodology introduced in this study can serve as a valuable tool for livestock managers to evaluate the nutrition of their dairy cattle and make informed decisions about their feeding practices.