Fuzzy modelling as an intelligent tool to study animal behaviour: An application to birds with environmental enrichment.

This study aimed to apply fuzzy modeling to the behavioral responses and physical activity of broiler chickens with and without environmental enrichment. 96 Cobb broiler chicks, raised in a controlled environment, 48 birds for each treatment (T1 = presence and T2 = absence of environmental enrichment) at different ages (1, 2, 3, 4, 5 and 6 weeks old) during the two production cycles were used. The behavioral assessment was carried out by recording the images for later analysis at the following observation times: 8–9 am, 12–1 pm and 4–5 pm, totaling three hours of observation per rearing week (1, 2, 3, 4, 5 and 6) at 5 min intervals of between image compositions. The experimental design was completely randomized in a split-plot, and the treatments and subplots were allocated to the weeks of rearing of the birds, containing 48 repetitions. The activity level of birds was based on behavioral data of broiler birds, in which fuzzy modeling was applied. Input variables (locomotion and activity) and output variables (activity level) of the fuzzy system were applied. During the first three weeks of life, the birds showed a moderate activity level, according to the fuzzy classification (between 0.300 and 0.450) between the two proposed treatments. At weeks 5 and 6, high levels of flock activity were characteristic of the group housed in an environment with the presence of environmental enrichment, while birds housed in the environment without enrichment showed a moderate level of activity at all times. At the times when high levels of activity were reported, the same behavior as in week 4 was observed: the activity level increased due to high values of "activity" behaviors, while "walking" behaviors remained in the moderate range. The fuzzy modeling revealed that the use of environmental enrichment (ladder with perch) increased the activity level of broilers in the last weeks, which displayed the highest levels of activity (mobility) of the birds. The use of fuzzy logic in behavioural responses allowed a greater global clarification of the birds' behavioural response pattern, while the traditional approach showed an increase in specific responses, which leads us to conclude that both approaches can be used synergistically. • Fuzzy modeling is recommended for studying animal behavior. • The application of fuzzy logic modeling made it possible to make a global inference on bird behavior, while the results of conventional analysis were specific to the individuality of behaviors. • The presence of environmental enrichment increased the activity level of chickens, especially in the last few weeks. [ABSTRACT FROM AUTHOR]