Rapid identification and quantification of vegetable oil adulteration in raw milk using a flash gas chromatography electronic nose combined with machine learning

A method for detecting vegetable oil adulteration in raw milk was established using a flash gas chromatography electronic nose (FGC E-nose) method combined with four machine-learning algorithms. Corn oil, palm oil (PO), soybean oil, and blended oil were added to skim milk samples to simulate adulteration. In the qualitative analysis, satisfactory results, with accuracies of 1.000 and 0.9565, and F1-scores of 0.9625 and 0.9778 were acquired using random forest (RF) and support vector machine models, respectively. In the quantitative analysis of PO concentration in adulterated raw milk, the RF algorithm achieved the best performance (R2 = 0.9792 and root mean square error = 0.2583) of the four algorithms tested in the prediction model. In model validation, practical inspection of actual samples verified the effectiveness of the proposed method at detecting vegetable oil adulteration in raw milk. In brief, an FGC E-nose method combined with machine learning is suggested to be an effective method for rapidly and precisely detecting vegetable oil adulteration in raw milk.