Suhandy, Diding (2018) THE PREDICTION OF SHELF LIFE OF LOCAL ORANGES USING SPECTRAL INFORMATION IN UV-VISIBLE-NIR REGION COMBINED WITH PARTIAL LEAST SQUARES REGRESSION. Malaysian Applied Biology (MAB), 47 (3). pp. 115-120. ISSN 0126/8643
|
Text
47_03_14 full paper.pdf Download (1MB) | Preview |
Abstract
Oranges is easy to be broken during handling and long transportation. One of the most challenging issues in this supply demand chain of oranges is to separate the fresh orange fruits from the older ones. During storage, the quantity of flavonoid substances in oranges is decreasing. In this research we investigate the potential application of using absorbance spectral information in UV-Vis-NIR region for prediction of shelf life in local orange fruits (Siam oranges from Jember) during storage. For this, we perform spectral acquisition of extracted orange samples in 1, 4, 7, 10 and 13 days of storages using a UV-Vis spectrometer in absorbance mode (GenesysTM 10S UV-Vis, Thermo Scientific, USA). For extraction samples we use a 2 cm x 2 cm of peel part of oranges. The sample preparation was done with chloroform as solvent for fluorescence substance extraction purpose. The calibration model for shelf life prediction of local oranges was developed using PLS regression with full cross validation. The calibration resulted in good correlation with r = 0.89 for calibration step and r = 0.63 for validation step, respectively. The prediction using different samples resulted in root mean square error of prediction (RMSEP) = 3.34 days. It can be concluded that there is a potential application of using spectral information in UV-Vis–NIR region combined with PLS regression for shelf life prediction of local oranges.
| Item Type: | Article |
|---|---|
| Subjects: | S Agriculture > S Agriculture (General) |
| Divisions: | Fakultas Pertanian (FP) > Prodi Teknik Pertanian |
| Depositing User: | Dr. Diding Suhandy |
| Date Deposited: | 14 Nov 2018 09:53 |
| Last Modified: | 14 Nov 2018 09:53 |
| URI: | http://repository.lppm.unila.ac.id/id/eprint/9920 |
Actions (login required)
 |
View Item |