Résumé:
This study aims to characterize local chicken farms in a specific region and to determine the socio-economic factors that can influence the hatching rate of eggs. To do this, an artificial neural network model was applied to predict the hatching rate from various socio-economic parameters, such as region, gender, age, level of education, experience (number of years), disinfection of premises and use of food supplements. The results showed that among the factors studied, the level of education was the most important to predict the hatching rate, followed by the age of the breeders. Other factors, such as region and experience, also contributed to the forecast, but to a lesser extent. The performance of the model was evaluated using various statistical criteria, with a correlation coefficient of 0.634 and an R² of 0.402, indicating a moderate correlation and the ability of a model to explain about 40% of the hatching variance. Although the results are promising, there is still room for improvement by adding more potential factors and optimizing the model hyperparameters. In conclusion, this study identified key factors influencing the hatching rate of local chickens and demonstrated the usefulness of artificial neural networks to predict agricultural outcomes based on socio-economic parameters. Future research could focus on including additional variables and improving modelling techniques to obtain even more accurate forecasts.
