Banca de DEFESA: HADJA LORENA RANGEL UCHOA CAVALCANTI DE MENEZES COSTA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : HADJA LORENA RANGEL UCHOA CAVALCANTI DE MENEZES COSTA
DATE: 28/02/2023
TIME: 09:00
LOCAL: Videoconfeência
TITLE:

ARTIFICIAL NEURAL NETWORKS: A PRELIMINARY STUDY FOR PREDICTION OF NUTRIENT INTAKE  AND PERFORMANCE OF HAIR SHEEP IN FEEDLOT SYSTEM

KEY WORDS:

Machine Learning, Artificial Intelligence, ANN, Santa Inês.

PAGES: 31
BIG AREA: Ciências Agrárias
AREA: Zootecnia
SUBÁREA: Nutrição e Alimentação Animal
SPECIALTY: Avaliação de Alimentos para Animais
SUMMARY:

As the main reason for confining meat animals is to regularly produce quality meat to supply the
domestic market with lower production costs and regular supply, the use of a system based on
Artificial Neural Networks (ANN) that allows performance prediction animal, based on initial entries
and historical series, looks promising, allowing low-performing animals to be identified in advance
and discarded in a way that maximizes production. Given the importance of this information, as well
as the consumption of nutrients by the animals, the present study aims to investigate the use of ANN
as predictors of nutrient consumption and, based on the chemical composition of the diet, predict the
performance of hair sheep in confinement . The database was obtained from a collection of scientific
experiments carried out in the sheep sector of the Department of Animal Science at the Federal Rural
University of Pernambuco. In order to test the effectiveness of the ANN regarding the prediction of the
desired characteristics, nine other prediction techniques with different characteristics were explored,
aiming at evaluating different scenarios and observing the best technique for each situation. For all
predicted features, such as average daily consumption, daily gain, and final body weight, ANN
performed far below expectations when compared to other techniques included in the analysis. This
may be a consequence of the low data volume available for training, verification and verification of
the prediction models. New studies should be carried out in order to explore more new prediction tools
that are promising for animal science as a whole.

 

COMMITTEE MEMBERS:
Presidente - ANTONIA SHERLANEA CHAVES VERAS
Externa ao Programa - 1660490 - RENATA VALERIA REGIS DE SOUSA GOMES - UFRPEExterno à Instituição - CLÁUDIO ELÍZIO CALAZANS CAMPELO - UFCG