Egg production curve fitting using least square support vector machines and nonlinear regression analysis

Abstract

It was aimed to model egg production curves using nonlinear regression analysis and least squares support vector machines in this study. The accuracy of the models was calculated using the Akaike information criteria, mean square error, mean absolute percentage error, mean absolute deviation, R-2 and AdjR(2). The data set consisted of egg performance values of laying hens recorded from 20 weeks to 70 weeks of age. The longitudinal data had a nonlinear structure. The results showed that the least squares support vector machines method, which is considered in different parameter combinations, can be used as an alternative to classical methods and predictions have lower errors. The present study shows that least squares support vector machine methods can be used successfully in the modelling of egg production curves in laying hens.