Changes and machine learning-based prediction in quality characteristics of sliced Korean cabbage (Brassica rapa L. pekinensis) kimchi: Combined effect of nano-foamed structure film packaging and subcooled storage

• Published in: Food science & technology Vol. 171; p. 114122
• Main Authors: Park, So Yoon, Kang, Miran, Yun, Suk-Min, Eun, Jong-Bang, Shin, Bo-Sung, Chun, Ho Hyun
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.12.2022
• Subjects: Headspace composition; Korean cabbage kimchi; Machine learning; Packaging; Storage conditions; R2; UV; MPE; RSC; AF; Headspace composition; RMSE; TA; LDPE; MAE; KCK; MLR; CFU; RF; Korean cabbage kimchi; Storage conditions; XGBoost; Machine learning; Packaging; NFS; PET; TLAB; ML
• ISSN: 0023-6438; 1096-1127
• DOI: 10.1016/j.lwt.2022.114122

The prevention of over-ripening during storage and distribution is crucial for commercial kimchi products with a short shelf life. This study evaluated the effects of nano-foamed structure (NFS) film packaging and subcooled storage on the characteristics and stability of sliced Korean cabbage (Brassica rapa L. pekinensis) kimchi (KCK). The experiments were performed as factorial tests based on completely randomized designs. Subcooled storage (−3 °C) delayed total lactic acid bacteria (TLAB) growth and suppressed the changes in pH, titratable acidity (TA), and reducing sugar content (RSC) of KCK compared to refrigerated storage (4 and 10 °C). The headspace CO2 concentrations in NFS-1 and NFS-2 film packaging were maintained at < 6.1% throughout the storage (36 d), irrespective of the storage conditions. These results indicate that NFS film suppresses CO2 accumulation wherein subcooled storage prevents the over-ripening of KCK. Furthermore, we combined machine learning (ML) algorithms with experimental results to predict the quality of KCK. The ML-based algorithms, random forest and extreme gradient boosting trees were more accurate in predicting TA, RSC, and TLAB counts than multivariable linear regression. The study demonstrates an improved packaging system and optimal storage conditions for KCK, which can be extended to other kimchi and fermented products. •A hurdle strategy for Korean cabbage kimchi (KCK) is reported.•Machine learning (ML) algorithms were used for predicting sliced KCK quality.•Subcooled storage delayed lactic acid bacteria growth and improved KCK quality.•Nano-foamed structure film packaging reduced package distension.•ML-based algorithms predicted the quality indices of KCK with high accuracy.