Near Infrared Technology for Determining Cacao Pod Husk Quality Attributes as Animal Feed by means of PLSR Approach

Abstract

Abstract Agro-industrial residues have been widely used as feedstuffs for animal production due to its abundant availability and relatively cheap. The aim of this study is to create a NIRS model in prediction of nutritive content’s cacao pod husk including dry matter, crude fat, and ash by means of partial least squares regression (PLSR) approach. This study utilized cacao pod husk acquisitioned by NIRS spectrum with the wavelength from 1000 to 2500 nm. Proximate analysis was applied to measure nutritive values of cacao pod husk. The result of the study indicated that NIRS technology by means of PLSR approach with the help of DT pretreatment can be used as sufficient model performance to predict nutritive values of cacao pod husk for crude fat, and ash with the value of 0.7 and 0.5 for r and R2 respectively. Meanwhile, the value for RMSEC was 0.3 and 0.7 and 1.5 for RPD. However, judging from prediction performance, current NIRS approach seems not to able to determine moisture content and dry matter due to lower R2 and r coefficient indexes. Thus, it required further spectra enhancement for a robust prediction. This study concluded that NIRS technology can be used as rapid and simultaneous model to predict nutritive values of feed samples.