Visible-NIR Spectroscopy and Least Square Support Vector Machines Regression for Determination of Vitamin C of Mandarin Fruit

摘要

A fast and non-destructive method was developed for determination of vitamin C of intact mandarin fruit by visible near infrared (visible-NIR) spectroscopy. A total of 69 samples were prepared for the calibration (n = 54) and prediction (n = 15) sets. The reflectance spectra of mandarin were obtained in the wavelength range from 450 to 1750 nm. The variables were selected for developing calibration models by partial least squares regression (PLSR) and least squares support vector machine (LS-SVM). The correlation coefficient for vitamin C was 0.83, and root mean square error of prediction (RMSEP) was 2.30 mg/100 g. The results were achieved when the variables of 519 nm, 548 nm, 608 nm, 676 nm, 680 nm, 1397 nm, 1401 nm, 1410 nm, 1475 nm and 1708 nm were utilized in conjunction with LS-SVM. This showed the capability of visible-NIR and the important role of chemometrics in developing accurate models for prediction vitamin C of intact mandarin fruit.