EGG QUALITY ASSESSMENT: A MODEL COMPARISON APPROACH USING BAYESIAN MIXED LOGIT, MIXED LOGIT, LOGISTIC REGRESSION AND MULTINOMIAL REGRESSION MODELS

This study compares the performance of Bayesian mixed logit, mixed logit, logistic regression, and multinomial regression models in analyzing egg quality. The results show that the Bayesian mixed logit model outperforms traditional models, with egg weights, shell thickness, and shape index emerging as significant determinants of egg quality. The Bayesian mixed logit model's superior performance is evident in its lower AIC, DIC, RMSE, and MAE values. These findings have implications for the poultry industry, highlighting the importance of considering complex relationships between egg quality traits.