Study on statistical characteristics of some non-linear functions in fitting egg production curve

The aim of this study was to determine the best predictive model for the accurate description of the average daily egg production being produced by the layers during the productive period. A total number of 81,565 daily egg records (1999-2000) were used. In the current study, Modified Gamma, Compartmental, Algebraic, Modified Compartmental, Logistic-Curvilinear, Narushin and Takma 2 (NT2) and Lokhorst functions were fitted using NLIN procedure of SAS software. Mean error sum of squares (MSE), adjusted coefficient of determination (Adj-R2), Akaike information criterion (AIC) and Durbin-Watson (DW) statistics were used for evaluation of the functions. The results revealed that Logistic-Curvilinear function was the most preferable one because its Adj-R2 value was the highest while it’s MSE and AIC values were lower than for the remaining functions. Assessing the goodness of fit of the used functions precisely, however, it was concluded that almost all of the functions are accurate enough in describing the average daily egg production curve. In such conditions, other features of functions should be considered and it seems that the simpler the function is mathematically the more desirable it is for the researcher or farm manager.