Classification of cow behavior using 3-DOF accelerometer and decision tree algorithm

Monitoring cattle motion is essential, since it helps farmers to take a comprehensive view of the cattle's health and estrus time. However, the issue is not able to supervise the cattle in a long time and raising many cattle especially. Therefore, this paper aim to build a device which can sense the states of behavior actives and researches a method to prognosticate the cattle's health by using a cattle monitoring device that can record the 3-axis acceleration to analyze. This sensor is used to measure three-axis accelerometer data on the natural behavior of Vietnamese Yellow cows that live in a cage. The data of the accelerometer output signal are used to modify a simple behavioral classification such as: lying, standing and feeding. Therefore, we can identify some of cattle health events like lameness and estrus cycle. The classification results were tested with the model of the cow.