Artificial Neural Network Model to Predict Crude Protein and Crude Fiber from Physical Properties of Feedstuffs

Abstract

The aim of this research was to build artificial neural networks model to predict crude protein and crude fiber content from physical properties of feedstuffs. The 91 data were obtained from https://repository.ipb.ac.id using keywords, e.g., sifat fisik and pakan. To reduce the dimensional of the data had been transformed. The independent variables consist of specific gravity (SG), bulk density (BD), compacted bulk density (CBD) and angle of repose (AoR). The dependent variable was crude protein (CP) and crude fiber (CF). Artificial neural networks (ANN) model built by R programing language 3.6.0 using library R-base and neuralnet. The correlation and accuracy used to compare predicted and actual. ANN model of crude fiber has an accuracy of 75.08% and Pearson's signification correlation (0.7529; P <0.01). ANN model of crude fiber has an accuracy of 75.08% and Pearson's signification correlation (0.7529; P <0.01). The artificial neural networks model generally can perform better to predict crude protein and crude fiber from physical properties of feedstuffs.