Application of electronic nose (E-nose) to estimate shelf-life of edible oils
學生姓名:
黃頌仁
指導教授:
孫寶年
學期:
111上
摘 要:
In the food industry, it is important that monitor edible oil quilty. However, A.O.A.C. and A.O.C.S. analytical methods have disadvantages such as solvent contamination, sample destruction and time consuming. It is necessary to develop a rapid and eco-friendly method for identifying the quality of edible oils. Generally, edible oils are affected by internal moisture, impurity, temperature and light during storage, which causes its oxidation products content to increase and smell changes. A portable metal oxide semiconductor gas sensor is one of the tools that have potential applications in food quality monitoring and agricultural management in recent years. Therefore, the aim of this study was to discuss the application of portable E-nose combined with machine learning in the prediction of edible oil storage period, and the potential of substituting traditional analysis methods. Stored edible oils were tested for order change by TGS822, TGS813, TGS2600, TGS2602, TGS2611, TGS2620, MQ3, MQ9, MQ135, MQ136 and MQ138, and combined with E-nose signal responses, Linear discriminant analysis (LDA), Quadratic discriminant analysis (QDA), k-Nearest neighbors (KNN), and Support vector machines (SVM) were compared to build a qualitative recognition model of storage time. The correct recognition rate of each model is greater than 80%, which is consistent with the results of A.O.A.C. methods. Overall, the E-nose array is effectively used in the monitoring of edible oil odor changes during storage. The E nose signal combined with the building of machine learning models, can not only effectively predict the storage period and quality of edible oil but also substitute traditional solvent analysis methods.
摘要.pdf