Estimating Shelf Life of Packed Fresh Milk Using Odor and Machine Learning: A Feasibility Study

This paper presents a feasibility study focused on estimating the shelf life of packed fresh milk through odor analysis and machine learning techniques. The study involved the capture of milk odors using five chemical gas sensors. Four distinct features were extracted from the sensor responses, namely response range, response ratio, initial response rate, and cumulative response. These features were then used as inputs for subsequent machine learning analysis. Storage tests were conducted on four brands of fresh milk, stored at room (25 °C) and refrigerated (10 °C) temperatures, resulting in the collection of features and remaining shelf-life data. A Multilayer Perceptron (MLP) neural network was used to develop predictors for both maximum and minimum remaining shelf life. The results indicated a high level of prediction accuracy, ranging from 85.0 to 100% for the prediction of maximum remaining shelf life. However, the minimum remaining shelf-life prediction exhibited lower accuracy, ranging from 43.1 to 67.6%. This discrepancy highlights the complexity in predicting the minimum shelf life, which relies solely on sensory analysis. This feasibility study presents a promising avenue for estimating the shelf life of food products by leveraging on odor data and machine learning algorithms. The findings underscore the potential of odor-based analysis as a complementary approach to traditional shelf-life estimation methods. However, the challenges associated with accurately predicting minimum shelf life based solely on odor require further investigation and refinement. In conclusion, this study offers valuable insights into the application of odor analysis and machine learning for shelf-life estimation, potentially enhancing food safety and quality assessment in the food industry. Further research and refinement of this methodology could lead to more accurate and practical shelf-life prediction techniques for a wide range of food products.