Astuti
Research Center for Processing Technology, National Research and Innovation Agency, South Tangerang, Banten, 15314, Indonesia

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Rapid Prediction of Total Gingerol Content and Harvest Age Classification of Fresh Indonesian Red Ginger Using UV-Vis Spectroscopy Tika Hafzara Siregar; Fareihan Rizki Farikhin; Slamet Widodo; Y. Aris Purwanto; Lamhot Parulian Manalu; Tri Yulni; Nenie Yustiningsih; Harianto; Jonni Firdaus; Primawati Yenni Fauziah; Endang Suwandi; Marga Asta Jaya Mulia; Astuti; Tantry Eko Putri Mariastuty
Science and Technology Indonesia Vol. 11 No. 4 (2026): October
Publisher : Research Center of Inorganic Materials and Coordination Complexes, FMIPA Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26554/sti.2026.11.4.1406-1422

Abstract

Red ginger (Zingiber officinale var. rubrum) is a potential spice commodity widely used in the food and pharmaceutical industries due to its high gingerol content, which has antioxidant, anti-inflammatory, and anticancer properties. This study aims to develop a predictive model for gingerol content and a classification model for harvest age in Indonesian fresh red ginger using chemometrics and machine learning based on Ultraviolet-Visible (UV-Vis) spectroscopy. Samples from three harvest ages (7, 8, and 9 months after planting) were analyzed using a UV-Vis spectrometer. UV-Vis spectral data were acquired and processed using several spectral preprocessing techniques, namely baseline correction, Savitzky-Golay (SGs), Multiplicative Scatter Correction (MSC), and first derivative (D1). The analysis was performed using Partial Least Squares (PLS) and Artificial Neural Network (ANN) for prediction, as well as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), k-NN, Random Forest, and Support Vector Machine (SVM) for harvest age classification. The ANN paired with SGs+D1 with 2 hidden nodes achieved the best predictive performance (R² = 0.66, r = 0.74), with RMSEC and RMSEP of 0.7062 and 0.8317, respectively. Meanwhile, the SVM model with a combination of Baseline+D1 preprocessing achieved a harvest age classification accuracy of 0.91 and precision, recall, and F1-score of 0.93 each. The findings of this research yield significant practical implications for both advanced post-harvest technology and the phytopharmaceutical industry. By delivering a highly precise classification model, this work lays a solid technical foundation required to design high-throughput, automated, and cost-effective sorting systems tailored for fresh Indonesian red ginger matrices.