Estimation and Prediction of Metabolizable Energy Contents of Wheat Bran for Poultry

The biological procedure used to determine the nitrogen-corrected True Metabolizable Energy (TMEn) value of feed ingredient is costly and time consuming. Therefore, it is necessary to find an alternative method to accurately estimate the TMEn content. In this study, ۲ methods of Multiple Linear Regression (MLR) and Artificial Neural Network (ANN) were developed to describe the TMEn (Kcal kg-۱ DM) value on a Dry Matter (DM) basis of Wheat Bran (WB) samples given their chemical composition of Ether Extract (EE), ash, Crude Protein (CP) and Crude Fiber (CF) contents (all used as % of DM). A data set containing ۱۰۰ WB samples were used to determine chemical composition and TMEn. Accuracy and precision of the developed models were evaluated given their produced prediction values. The results revealed that the developed ANN model [R۲= ۰.۹۰; Root Mean Square Error (RMSE)= ۶۴.۰۷ Kcal kg-۱ DM for training set; and R۲= ۰.۸۹; RMSE= ۸۲.۶۹ Kcal kg-۱ DM for testing set] produced relatively better prediction values of TMEn in WB than those produced by conventional MLR [R۲= ۰.۸۱; RMSE= ۸۶.۷۶ Kcal kg-۱ DM for training set; and R۲= ۰.۸۴; RMSE= ۸۶.۶۱ Kcal kg-۱ DM for testing set]. The developed ANN model may be considered as a promising tool for modeling the relationship between chemical composition and energy of WB samples. To provide the users with an easy and rapid tool, an Excel® calculator, namely, ANN_WB_ME_Poultry, was created to predict the TMEn values in WB sample given its chemical composition and using the developed ANN model.

کلیدواژه ها:

Metabolizable energy ، Prediction model ، Wheat bran.