Classification of crop flours based on protein contents using near infra-red spectroscopy and principle component analysis. (November 2019)

• Record Type: Journal Article
• Authors: Yuliyanda, I; Masithoh, R E; Khuriyati, N; Saputro, A D
• Is Part Of: IOP conference series. Volume 355(2019)
• Publication Date: 2019-11
• Subjects: Earth sciences -- Periodicals; Environmental sciences -- Congresses; Environmental sciences -- Periodicals; 550.5
• DOI: 10.1088/1755-1315/355/1/012002
• Languages: English
• ISSNs: 1755-1307

Abstract

Abstract: Indonesia has abundant crops that can be used as carbohydrate sources. Those crops are made into flour to prolong the shelf life, as well to ease for handling and cooking. Crops as carbohydrate sources are usually high energy but low protein. The aim of this research was to classify flours made of various crops using near infra-red spectroscopy (NIRS) and principle component analysis (PCA). The samples used in this study were six types of flour made of banana, breadfruit, taro, arrowroot, purple sweet potato, and modified cassava (mocaf). The reflectance data were taken using the NIRFlex N500 Fiber Optic Solids Cell at wavelengths of 1000-2500 nm. The spectral obtained were pre-processed and analyzed using The Unscrambler X version 10.5.1. Three pre-processing methods were used, i.e. 1st Savitzky Golay Derivative, Normalization, and Standard Normal Variate (SNV). PCA was able to classify flours based on types of crops. The best transformation was SNV which was able to classify all groups of samples with 100% success rate. PCA model was also able to differentiate low and high protein level of samples aligned with the chemical analysis.