Classification and Prediction of Chilli Powder (Capsicum Annuum L.) Colour Based on Near-Infrared Spectra and Random Forest
Pertanika Journal of Tropical Agricultural Science · 30.10.2026 · 49 (5)
DOI: 10.47836/pjtas.49.5.23 ↗Kurz gefasst
Fachartikel in der Zeitschrift Pertanika Journal of Tropical Agricultural Science, online erschienen am 30.10.2026. Verfasst von Nafis Khuriyati und 4 weiteren Autor:innen.
Abstract (Originalsprache)
Colour measurements play a crucial role in the chilli industry for establishing colour standards. Limited research has been conducted on predicting colour characteristics using Near Infrared (NIR) measurements, yet none of them has been applied to chilli powder. This study aimed to classify and predict the colour intensity of chilli powder using chemometric models based on NIR spectroscopy and the random forest algorithm. This study used 84 chilli powder samples from four colour groups. NIR spectra were measured with NIRFlex N-500 at 1000 to 2500 nm. Three pretreatments were applied: Multiplicative Scatter Correction (MSC), first derivative Savitzky-Golay, and detrending. Random forest was used for classification and prediction of L*, a*, and b*. Data were split into 75% training and 25% testing, evaluated with accuracy score and root mean square error (RMSE). The machine learning method of random forest has been successfully applied for classifying and predicting the colour of chilli powder based on NIR spectroscopy. Classification of four colour groups achieved the best accuracy of 0.9524 with first derivative Savitzky-Golay, while raw spectra and MSC reached 0.7619. L* was predicted most accurately with MSC (RMSE 2.68), meanwhile a* and b* were predicted most accurately with first derivative Savitzky-Golay (RMSE 2.86 and 3.0821). The findings indicate that the most effective pretreatment varied according to colour index.
Zitieren
APA
Khuriyati, N., Sukartiko, A. C., Alfiani, R. N., Ahmadi, T. P., & Kasih, A. N. (2026). Classification and Prediction of Chilli Powder (Capsicum Annuum L.) Colour Based on Near-Infrared Spectra and Random Forest. Pertanika Journal of Tropical Agricultural Science, 49(5). https://doi.org/10.47836/pjtas.49.5.23
Vancouver
Khuriyati N, Sukartiko AC, Alfiani RN, Ahmadi TP, Kasih AN. Classification and Prediction of Chilli Powder (Capsicum Annuum L.) Colour Based on Near-Infrared Spectra and Random Forest. Pertanika Journal of Tropical Agricultural Science. 2026;49(5). doi:10.47836/pjtas.49.5.23
BibTeX
@article{Khuriyati2026,
author = {Khuriyati, Nafis and Sukartiko, Anggoro Cahyo and Alfiani, Ririn Nur and Ahmadi, Tyasto Prima and Kasih, Avicenna Nur},
title = {Classification and Prediction of Chilli Powder (Capsicum Annuum L.) Colour Based on Near-Infrared Spectra and Random Forest},
journal = {Pertanika Journal of Tropical Agricultural Science},
year = {2026},
volume = {49},
number = {5},
doi = {10.47836/pjtas.49.5.23}
}
EUVIMED