Comparative analysis of classical growth models and artificial neural networks in predicting egg production parameters in three commercial broiler parent stocks.

In the rapidly evolving poultry industry, effective management is dependent on precise forecasting of production metrics. As producers seek to maximize yield while minimizing costs, the ability to accurately predict egg production parameters becomes essential. Traditional growth models have long been employed for this purpose, but with advancements in computational techniques, artificial neural networks (ANNs) present a promising alternative. This paper examines the efficacy of various classical growth models, Gamma, compartmental, logistic, curvilinear Gompertz, Richards, and Morgan models, alongside a multi-layer feed-forward ANN model in predicting weekly and cumulative production metrics of three distinct commercial broiler parent stock (Ross, Arbor Acres, and Cobb). The analysis revealed that the Morgan model performed as the most accurate for modeling cumulative data, showcasing a superior fit characterized by high adjusted R² values and low RMSE, AIC, and BIC. In contrast, the Logistic - curvilinear model demonstrated high precision for weekly data. The ANN significantly surpassed the best traditional Logistic curvilinear and Morgan models in predicting weekly and cumulative production metrics based on RMSE, AIC, and BIC criteria (RMSE = 100% both for weekly and cumulative productive metrics, AIC = 91.6% for weekly and 58.3% cumulative productive metrics, and BIC = 66.6% both for weekly and cumulative productive metrics). Although the findings underscore the strength of the classical growth model for cumulative data forecasts in poultry production systems, the advantages of ANNs in the prediction of egg production metrics represent a vital advancement toward more responsive flock management strategies. By integrating classical growth models with ANN approaches, poultry producers can enhance decision-making processes, leading to improved productivity and sustainability. The integration of these methodologies provides a powerful framework for accurate forecasting, ultimately advancing the poultry industry's capacity for efficient production management.