Partial Least Square Regression Prediction Model for Strawberry Freshness Monitoring During Storage Using Electronic Nose

Strawberries produce undesirable aromas because of chemical reactions and microbial growth during deterioration. An electronic nose can identify the changes in these aromas. This study aimed to analyze the quality indexes and volatile organic compounds (VOCs) profile changes in strawberries at constant temperatures during a specific storage period. Correlation and principal component analyses (PCAs) were also employed to determine the partial least squares regression (PLSR) for predicting the screened quality indexes using the VOC data. The results showed a high correlation of quality indexes as a function of time, including weight loss, total yeast and mold growth, total color differences, and texture (firmness). PCA revealed the contribution of VOCs in the sample, suggesting that propyl nonanoate, linalool, 2-propanol, and (Z)-3-hexen-1-ol were freshness indicators of strawberries. PLSR appropriately predicted weight loss, total yeast and mold growth, total color differences, and texture (firmness) between the observed and predicted data, which were expressed as R² = 81%–96%, RMSE = 21%–97%, and MAE = 15%–75%; however, the model with nonlinearity performed poorly. This finding provides important information for future development in the storage of strawberries during distribution from farm to retail at other constant temperatures or in fluctuating temperatures.