Software based on fuzzy logic for the classification of body mass cattle

• Published in: Soft computing (Berlin, Germany) Vol. 28; no. 13-14; pp. 8151 - 8165
• Main Authors: Maziero, Luana Possari, Chacur, Marcelo George Mungai, Cremasco, Camila Pires, Bordin, Deyver, Putti, Fernando Ferrari, Gabriel Filho, Luís Roberto Almeida
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.07.2024; Springer Nature B.V
• Subjects: Agriculture; Animals; Application of Soft Computing; Artificial Intelligence; Automation; Body mass index; Body size; Cameras; Cattle; Chronology; Classification; Computational Intelligence; Control; Decision making; Engineering; Farms; Fuzzy logic; Fuzzy sets; Fuzzy systems; Management decisions; Mathematical Logic and Foundations; Mechatronics; Methods; Neural networks; Production costs; Robotics; Set theory; Software; Variables; Fuzzy logic; Body mass index; Score; Artificial intelligence
• ISSN: 1432-7643; 1433-7479
• DOI: 10.1007/s00500-024-09699-8

The body mass index (BMI) is used by farmers to determine the body condition of animals, either to determine the best moment for slaughter or to classify animals for minimizing production costs. The objective of the present work was to establish the efficiency of a software based on fuzzy artificial intelligence systems capable of evaluating several herds of different breeds, comparing them and providing rural producers and researchers with additional parameters for decision making on body mass index. For this, systems based on fuzzy rules were created, with “mass” and “height” as input variables, and a new BMI called the Fuzzy BMI as the output variable. The output variable was classified into five groups: “Very Low”, “Low”, “Medium”, “High,” and “Very High”, allowing farmers to compare animals. A software program was developed in the Delphi language, and the user interface was tailored for cattle breeders, enabling them to evaluate their herds. For model validation, a database of three farms was used, which totaled 287 animals of different ages and breeds. The developed models presented a Pearson correlation greater than 0.92, a value much higher than the traditional models. According to the Fuzzy BMI determined for each animal and the combinations between all animal masses and heights, the developed software allows farmers to enter data and obtain answers that can be used to improve management, i.e., optimize resources and increase profit.