Aplicación para el análisis de las etapas de desarrollo embrionario de gallus gallus domesticus basado en redes neuronales artificiales

Abstract

Small-scale farms are places where not only animals are raised, but certain species are also bred, in this case focusing on poultry, therefore, egg incubations are carried out. The present project aims to analyze the stages of embryonic development of gallus gallus domesticus using artificial neural networks for more efficient control and monitoring through a web application, focusing on domestic breeding farms located in the city of Machala, province of El Oro, considering the great importance it has within small-scale farms that practice family poultry farming; therefore, it would be extremely beneficial for small businesses to have this valuable tool to develop appropriate strategies before and during incubation. The study developed has a quantitative, descriptive approach with a quasi-experimental design, carried out with a population of approximately 24,000 images corresponding to embryonic egg development. For the construction of the project prototype, two different methodologies were applied, the extreme programming (XP) and CRISP-DM methodology. The tests yielded both favorable and unfavorable results, with the system's relationship to the real world having a 30.8% positive critique, which needs improvement, although the application favors compliance with most usability principles. On the other hand, the initial stage accepts the hypothesis with an 82% precision, and in the remaining cases, the hypothesis is rejected with precisions of 73%, 70%, and 55%, respectively. The results can be improved by correcting the overfitting of the neural network model, as well as obtaining images from the use of the application.