Classifier’s Performance for Detecting the Pecking Pattern of Broilers during Feeding

Broiler feeding is an efficient way of evaluating growth performance, health, and welfare status. This assessment might include the number meals, meal period, ingestion rate, intervals, proportion time spent eating. These parameters can be predicted by studying birds’ pecking activity. The present study aims to design, examine, validate classifying algorithms determine individual bird patterns at feeder. Broilers were reared from 1 42 days, with feed water provided ad libitum. A feeder equipped a force sensor was installed used birds starting 35 days age, acquire data during until days. obtained organized into two datasets. first comprises 17 attributes, supervised attribute ‘pecking detection’ classes, ‘non-pecking’ ‘pecking’ analyze classifiers. In second dataset, ‘maximum value’ discretized in three classes compose new dataset comprising classes’ non-pecking, light pecking, medium, strong. We developed validated models broiler classifiers (KNN, SVM, ANN) achieved high accuracy, greater than 97%, similar results all investigated scenarios, proving capable performing task detecting pecking.