Goat Feeding Behavior Modeling and Monitoring Utilizing Livestock Wearable Device and Machine Learning Algorithm

Goat farming shows potential as one of the so-lutions for providing quality livestock products in developing countries. Goat health is essential in determining the quality of goods obtained from the goat. Smart livestock farming has been a reliable solution to achieve high-yield and high-quality goat products. However, current smart livestock monitoring systems do not consider the feeding pattern and behavior of the animal as a factor in its overall health. Hence, this study aims to develop a system that effectively monitors the feeding behavior of the goat using wearable devices, machine learning, and a visualization interface. The wearable device consists of an accelerometer, gyrometer, and temperature sensors to monitor the goat's feeding behavior and movement. The data collected by the wearable device are wirelessly transmitted to a base station to build a sensor-based dataset. This dataset is used to generate a k-means clustering model to classify the feeding behavior of the goats. Results show that the system was able to classify the feeding behavior of the goat with an accuracy of 98.02 % and a reliability of 100% in comparison with the actual goat behavior on the farm. Lastly, a graphical user interface is developed to provide monitoring and other relevant information to help the farmers in grazing the goats.