Optical characterization of NIR spectra for chemomectric model of cocoa pod husk fermented for animal feed

Abstract

Cocoa pods husk (CPH) fermented can be used as an alternative animal feedstuff. Near-infrared (NIR) spectroscopy has been shown in order to determine the nutritional contents of CPH fermented. This study aims to assess the characteristics of the NIR raw spectra and pretreated spectra for chemomectric model. The raw spectrum pretreatments used include multiplicative scatter correction (MSC), Savitzky-Golay smoothing (SG), and first derivative (1st D) and a combination of each pretreatment, namely MSC + SG, MSC + 1st D, SG + 1st D, and MSC + SG + 1st D. The results showed that the NIR spectrum of CPH fermented had six absorption peaks associated with CPH nutrient content such as moisture (1450 and 1940 nm), lipids (1200 and 1731 nm), starch (2380 nm), and protein (2125 nm). MSC tries to eliminate the scattering and correct the differences in the baseline and trend. SG smoothing is used to remove high-frequency random noise, and peaks. 1st D can be used for baseline correction. Combination of each pretreatment, namely MSC + SG, MSC + 1st D, SG + 1st D, and MSC + SG + 1st D. The combination of pretreatment methods can reduce scattering and noise interference such as offset, dispersion, and overlap and improve smoothness. In addition, information relating to the nutritional contents of CPH was clearly highlighted. This study concludes that a combination of pretreatment methods is better at revealing hidden information and reducing noise than a specific pretreatment method. Combination of MSC + SG pretreatment in this study is considered to generate a spectrum that has good characteristics to be used in chemometric modeling due to it is able to provide information regarding the nutritional contents of CPH clearly, both for qualification and quantification.