In International journal of biometeorology
In this study, we develop an artificial intelligence model to predict the vulnerability of broiler production systems (broilers and facilities) to heat conditions using a fuzzy model approach. The model was designed with a multiple-input and a single-output (MISO) approach (input: physical environment and broilers age; output: degree of vulnerability of broilers system). For the validation of the fuzzy model, two approaches were used: (1) records from the scientific literature and (2) meteorological forecasts. First, we validated the model fuzzy with data from the scientific literature; second, we validate the model with data from meteorological forecasts. Both validation approaches were performed in different scenarios of the thermal environment (comfort, discomfort, and discomfort + low heat exchange), broilers' age (21-35 days, 25-39 days, and 28-42 days), and relative cooling efficiency (0% inefficient; and 80% efficient). Then, we applied the model to predict the degree of vulnerability of the broiler system with the help of weather forecasts. The recall and precision of the fuzzy model were high (> 0.9) for the thermal comfort and thermal discomfort + low heat exchange scenarios. In contrast, the fuzzy model was moderate agreement (recall 0.45; precision 0.64) for the thermal discomfort scenario compared to the scientific literature. The application of the model with the weather forecast showed the interaction between the physical and biological systems when submitted to a thermal environment challenge. Regardless of the broilers' age, a high degree of vulnerability was observed in facilities with inefficient cooling system. The fuzzy model developed in this study was efficient to predict the vulnerability of the broiler production system to heat conditions, further, to identify the uncertain conditions associated with broilers' age, relative humidity, and the relative cooling efficiency of the facilities.
De-Sousa Karolini Tenffen, Deniz Matheus, Santos Maurício Portella Dos, Klein Daniela Regina, Vale Marcos Martinez do
2023-Jan-28
Artificial intelligence, Decision-making, Expert system, Precision livestock farming
