PREDIKSI VITAMIN C, TOTAL ASAM TERTITRASI, DAN TOTAL PADATAN TERLARUT PADA BUAH MANGGA MENGGUNAKAN NEAR-INFRARED REFLECTANCE SPECTROSCOPY

Abstract

ABSTRAK Buah mangga memiliki kandungan nutrisi yang bermanfaat bagi kesehatan. Komoditas ini tersedia di pasar dalam berbagai jenis kulitvar, diantaranya cengkir, kweni, kent dan palmer. Tujuan penelitian ini adalah memprediksi kualitas internal buah mangga yaitu vitamin C, total asam tertitrasi, total padatan terlarut (TPT) menggunakan near-infrared reflectance spectroscopy (NIRS). Spektra diffuse reflectance yang diperoleh dari hasil akuisisi spektra dikoreksi dengan pra-pengolahan spektra metode orthogonal signal correction (OSC), dan turunan pertama savitzky-golay (dg1). Hasil penelitian menjelaskan bahwa vitamin C, total asam tertitrasi, dan TPT dapat diprediksi dengan baik menggunakan NIRS. Pra-pengolahan spektra memberikan pengaruh terhadap akurasi pendugaan kualitas internal buah mangga. Spektra dg1 memperoleh nilai R2 kalibrasi tertinggi pada ketiga parameter kualitas yaitu 0,98 (vitamin C), 0,87 (total asam tertitrasi), dan 0,96 (TPT). Namun, pada pendugaan vitamin C dan total padatan terlarut, spektra dg1 menampilkan nilai konsisten yang rendah yaitu sebesar 56% dan 63%. Pra-pengolahan spektra OSC mampu mereduksi jumlah faktor pada spektra original. Model kalibrasi terbaik pendugaan vitamin C dan total padatan terlarut diperoleh oleh spektra original, sedangkan pendugaan total asam tertitrasi ditunjukkan oleh spektra dg1 Kata kunci: Analisis Multivariat; Kalibrasi; Kemometrika; Nondestruktif; Validasi ABSTRACT Mango fruit contains a lot of beneficial nutrition for health. This commodity is available at the market in various cultivars, including cengkir, kweni, kent, and palmer. The purpose of the research was to predict the internal quality of mango such as vitamin C, titratable acidity, soluble solid content using near-infrared reflectance spectroscopy (NIRS). Diffuse reflectance spectra acquired from spectra acquisition were corrected using spectra pre-processing methods of orthogonal signal correction (OSC) and first derivative savitzky-golay (dg1). The results explained that vitamin C, titratable acidity, and SSC were able to be properly predicted using NIRS. Spectra pre-processing gave effect to the accuracy of internal quality prediction of mango. Dg1 spectra obtained the highest calibration R2 in the three quality parameters of 0.98 (vitamin c), 0.87 (titratable acidity), and 0.96 (SSC). However, in vitamin c and SSC prediction, dg1 spectra yielded low consistent values of 56% and 63%. Besides, OSC spectra pre-processing was able to reduce the number of factors in the original spectra. The best calibration model for predicting vitamin c and total dissolved solids was achieved by the original spectra, while the prediction of titratable acidity was shown by dg1 spectra Keywords: Calibration; Chemometrics; Multivariate Analysis; Nondestructive; Validation