Tomato Shelf Life Prediction using IoT and Machine Learning

Predicting tomato shelf life is crucial for optimizing supply chains and reducing waste. This study utilizes IoT and machine learning to analyze postharvest characteristics of tomatoes in four ripening stages under controlled conditions until the rotten phase. Key findings include significant reductions in weight and dimensions, increased red and decreased green color components indicating ripening, and longer shelf life for unripe tomatoes. Green 1 tomatoes had the most extended shelf life of 18 days. $\mathbf{p H}$ trends showed higher levels during ripening (pH 3.5) compared to the green stage ($\mathbf{p H} 4.5$). To predict the shelf life, various machine learning models were applied and the Random Forest Regression was the most accurate model, exhibiting the lowest errors and highest predictive reliability.