Analysis the relationship of physiological, environmental, and cow milk productivity using AI

Abstract

The dairy cattle productivity is very depending on the quality of their environment and physiological aspect. Hence, the purpose of the paper is to looking for the relationship model of physiological, environmental and milk productivity by using artificial intelligence (AI). The model will be useful for the user to decide the best cow treatment in order to gain the best milk production. The research is started with literature review and early survey of cattle physiological, environment factors and milk productivity. The next step is measuring the environment data (temperature, wind speed, and relative humidity) and measuring physiological aspect (heart rate, body temperature) correlated with milk productivity in 500 pairs of data. All the data are collected and stored into the database and then trained and validated using Back Propagation Neural Network (BPNN) with Genetic Algorithm (GA) optimization. The initial BPNN architectures are selected in 2 hidden layer, delta bar delta learning rule, sigmoid transfer function and epoch 10000. As a result, the system successfully developed an intelligent tool to predict milk production in any levels of environment and physical condition. Based on sensitivity analysis, the relative humidity, heart rate, environment and cow body temperature are categorized in strong impact, beside that are in weak impact on milk production.