All Issue

2025 Vol.70, Issue 4 Preview Page

Original Research Article

Machine Learning Analysis of Rice Quality and Palatability Evaluation Data to Predict Rice Taste and Interpret Variables

쌀 품질 및 기호도 평가 데이터의 머신러닝 분석을 통한 밥맛 예측 및 변수 해석

Yuchan Choi1*
,
Hyun-jin Park1
,
Yu-geun Oh1
,
Ji-eun Kwak2
최 유찬1*
,
박 현진1
,
오 유근1
,
곽 지은2
1Junior Scientist, Quality Management and Evaluation Research Division, Department of Food Sciences, National Institute of Crop Science and Food, Rural Development Administration, Suwon 16613, Korea
2Junior Scientist, Planning and Coordination Division, National Institute of Crop Science and Food, Rural Development Administration, Wanju 55365, Korea
1국립식량과학원 식품자원개발부 품질관리평가과 농업 연구사
2국립식량과학원 기획조정과 농업 연구사
*Corresponding Author
1 December 2025. pp. 222-233
PDF XML Citation
Abstract

This study aimed to develop a machine learning-based evaluation framework for predicting rice (Oryza sativa L.) eating quality using physicochemical and sensory datasets accumulated over the past decade at the National Institute of Crop Science (NICS). A total of 297 rice samples were collected, including grain shape, milling, and physicochemical traits, along with sensory panel scores of overall palatability. After data preprocessing and the removal of outliers, 216 samples were used for model training. Principal component analysis (PCA), cluster analysis, and preference mapping were conducted to identify the structural relationships among rice quality traits. Gradient Boosting Machine (GBM) and Random Forest models were then trained to classify overall eating quality into two categories (A: superior, B: inferior). Model performance was evaluated using five-fold cross-validation and an independent test set. The Random Forest model achieved the highest accuracy (0.7143) and Kappa coefficient (0.4273), indicating statistically significant generalization performance (p = 0.009). SHAP (SHapley Additive exPlanations) analysis revealed that amylose, protein, and Milled/Brown Ratio were the most influential predictors of eating quality, where lower amylose and protein content and higher Milled/Brown Ratio contributed positively to superior eating quality (A). These results demonstrate the potential of machine learning as a complementary approach to sensory testing by enhancing objectivity and efficiency, and suggest core quality indices for rice breeding and processing—namely, amylose, protein, and Milled/Brown Ratio.

Keywords
amylose
machine learning
milled/brown ratio
preference mapping
protein
random forest
rice eating quality
SHAP
References
1

Bagheri, R. 2022. Introduction to shap values and their application in machine learning. Medium. 8th Aug.

2

Bhatnagar, S. R., Y. Yang, B. Khundrakpam, A. C. Evans, M. Blanchette, L. Bouchard, and C. M. Greenwood. 2018. An analytic approach for interpretable predictive models in high‐dimensional data in the presence of interactions with exposures. Genet. Epidemiol. 42(3) : 233-249.

10.1002/gepi.2211229423954PMC6175336
3

Breiman, L. 2001. Random forests. Mach. Learn. 45(1) : 5-32.

10.1023/A:1010933404324
4

Cao, N., W. Zhou, F. Zhao, G. Jiao, L. Xie, A. Lu, J. Wu, M. Zhu, Y. Liu, J. Yu, R. Zhao, X. Yang, S. Hu, Z. Sheng, X. Wei, Y. Lv, S. Tang, G. Shao, and P. Hu. 2025. OsGATA7 and SMOS1 cooperatively determine rice taste quality by repressing OsGluA2 expression and protein biosynthesis. Nat. Commun. 16(1) : 3513.

10.1038/s41467-025-58823-140223143PMC11994747
5

Chan, J. Y. L., S. M. H. Leow, K. T. Bea, W. K. Cheng, S. W. Phoong, Z. W. Hong, and Y. L. Chen. 2022. Mitigating the multicollinearity problem and its machine learning approach: a review. Mathematics 10(8) : 1283.

10.3390/math10081283
6

Cuili, W., G. Wen, H. Peisong, W. Xiangjin, T. Shaoqing, and J. Guiai. 2022. Differences of physicochemical properties between chalky and translucent parts of rice grains. Rice Sci. 29(6) : 577-588.

10.1016/j.rsci.2022.03.002
7

Faye, P., D. Brémaud, E. Teillet, P. Courcoux, A. Giboreau, and H. Nicod. 2006. An alternative to external preference mapping based on consumer perceptive mapping. Food Qual. Preference 17(7-8) : 604-614.

10.1016/j.foodqual.2006.05.006
8

Gong, X., L. Zhu, A. Wang, H. Xi, M. Nie, Z. Chen, Y. He, Y. Tian, F. Wang, and L. Tong. 2022. Understanding the palatability, flavor, starch functional properties and storability of indica-japonica hybrid rice. Molecules 27(13) : 4009.

10.3390/molecules2713400935807256PMC9268750
9

Greenwell, B. M., B. C. Boehmke, and A. J. McCarthy. 2018. A simple and effective model-based variable importance measure [Preprint]. arXiv. https://arxiv.org/abs/1805.04755

10.32614/CRAN.package.vip
10

Ishwaran, H. 2007. Variable Importance in Binary Regression Trees and Forests. Electron. J. Stat. 1 : 519-537.

10.1214/07-EJS039
11

Kim, C. S., N. Kim, and K. Y. Kwahk. 2019. Research trends analysis of machine learning and deep learning: Focused on the topic modeling. J. Korea Soc. Digit. Ind. Inf. Manag. 15(2) : 19-28.

12

Kumar, A. 2022. Machine Learning-Sensitivity vs Specificity Difference. Machine Learning, Data Analytics, date access, pp. 1-6.

13

Li, C., S. Yao, B. Song, L. Zhao, B. Hou, Y. Zhang, F. Zhang, and X. Qi. 2023. Evaluation of cooked Rice for eating quality and its components in Geng Rice. Foods 12(17) : 3267.

10.3390/foods1217326737685200PMC10486766
14

Liu, Q., Y. Tao, S. Cheng, L. Zhou, J. Tian, Z. Xing, G. Liu, H. Wei, and H. Zhang. 2020. Relating amylose and protein contents to eating quality in 105 varieties of Japonica rice. Cereal Chem. 97(6) : 1303-1312.

10.1002/cche.10358
15

Lundberg, S. M. and S. I. Lee. 2017. A unified approach to interpreting model predictions. Advances in neural information processing systems, 30.

16

Mohammed, A. and R. Kora. 2023. A comprehensive review on ensemble deep learning: Opportunities and challenges. J. King Saud Univ-Comput. & Inf. Sci. 35(2) : 757-774.

10.1016/j.jksuci.2023.01.014
17

Pudjihartono, N., T. Fadason, A. W. Kempa-Liehr, and J. M. O’Sullivan. 2022. A review of feature selection methods for machine learning-based disease risk prediction. Front. Bioinf. 2 : 927312.

10.3389/fbinf.2022.92731236304293PMC9580915
18

Schreurs, M., S. Piampongsant, M. Roncoroni, L. Cool, B. Herrera-Malaver, C. Vanderaa, F. A. Theßeling, Ł. Kreft, A. Botzki, P. Malcorps, L. Daenen, T. Wenseleers, and K. J. Verstrepen. 2024. Predicting and improving complex beer flavor through machine learning. Nat. Commun. 15(1) : 2368.

10.1038/s41467-024-46346-038531860PMC10966102
19

Shapley, L. S. 1951. Notes on the n-person game—II: the value of an n-person game.

20

Youn, Y. and Y. S. Kim. 2015. Physicochemical properties of rice varieties for manufacturing frozen fried rice. Food Sci. Preserv. 22(6) : 823-830.

10.11002/kjfp.2015.22.6.823
21

Zhou, L., S. Pan, J. Wang, and A. V. Vasilakos. 2017. Machine learning on big data: Opportunities and challenges. Neurocomputing 237 : 350-361.

10.1016/j.neucom.2017.01.026
Information
  • Publisher :The Korean Society of Crop Science
  • Publisher(Ko) :한국작물학회
  • Journal Title :The Korean Journal of Crop Science
  • Journal Title(Ko) :한국작물학회지
  • Volume : 70
  • No :4
  • Pages :222-233
  • Received Date : 2025-11-05
  • Revised Date : 2025-11-15
  • Accepted Date : 2025-11-18
  • DOI :https://doi.org/10.7740/kjcs.2025.70.4.222
Journal Informaiton The Korean Journal of Crop Science The Korean Journal of Crop Science
Online Submission
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close