AMEn Predictor: A mobile app to predict energy values of broilers feedstuffs

Abstract

This application note introduces an app for Android smartphones to predict the nitrogen-corrected metabolizable energy (AMEn) values of feedstuffs for broilers by using an artificial neural network model. This application, called AMEn Predictor, has been developed for Android devices. To calculate the AMEn prediction, the user has to insert values of the chemical composition of one concentrate feedstuff sample and the dry matter value to consider in the prediction. The dataset contains 48 bioassay results used to test the validation and efficiency of the tool. All predictions were processed instantly, with no internet connection required. Findings indicated a low value of mean absolute percentage error (MAPE = 5.92%) in the test dataset. The results confirm the fact that AMEn Predictor enables the animal nutritionists to obtain quick and accurate predictions of the AMEn for feedstuff samples, generally used by the poultry feed industry. AMEn Predictor was the first app for the prediction of broilers’ feedstuffs energy values currently freely available on Google Play.