Pattern recognition based on machine learning identifies oil adulteration and edible oil mixtures

Abstract

Previous studies have shown that each edible oil type has its own characteristic fatty acid profile; however, no method has yet been described allowing the identification of oil types simply based on this characteristic. Moreover, the fatty acid profile of a specific oil type can be mimicked by a mixture of 2 or more oil types. This has led to fraudulent oil adulteration and intentional mislabeling of edible oils threatening food safety and endangering public health. Here, we present a machine learning method to uncover fatty acid patterns discriminative for ten different plant oil types and their intra-variability. We also describe a supervised end-to-end learning method that can be generalized to oil composition of any given mixtures. Trained on a large number of simulated oil mixtures, independent test dataset validation demonstrates that the model has a 50th percentile absolute error between 1.4–1.8% and a 90th percentile error of 4–5.4% for any 3-way mixtures of the ten oil types. The deep learning model can also be further refined with on-line training. Because oil-producing plants have diverse geographical origins and hence slightly varying fatty acid profiles, an online-training method provides also a way to capture useful knowledge presently unavailable. Our method allows the ability to control product quality, determining the fair price of purchased oils and in-turn allowing health-conscious consumers the future of accurate labeling.

Rapid Identification of Adulteration in Edible Vegetable Oils Based on Low-Field Nuclear Magnetic Resonance Relaxation Fingerprints

“…These studies have demonstrated the substantial potential for applying LF-NMR relaxometry for the rapid identification of oil adulteration. However, subsequent research has demonstrated that the identification capability of conventional LF-NMR relaxometry is often limited by the fact that the relaxation properties of individual edible oils can be very similar to those of complex mixtures including two or more edible oils [ 27 ]. As a result, the relaxation properties of edible oils captured by conventional LF-NMR spectroscopy may lack the discrimination capability required for meeting the needs of public safety and business interests.…”

Discrimination, Quantitation, and Identification of Edible Vegetable Oil Blends Based on Their Fatty Acid Profiles

“…As the composition of crude oil is highly variable, depending upon the plant species, geographical location of the source and method of oil extraction [28]. Additionally, unintentional or intentional use of cheaper (i.e., lower quality), less pure, or mislabeled raw materials can occur during the oil blending or extraction process [29,30]. Addition of unwanted additives or mislabeling (especially if a manufacturer produce several type of oil blends in parallel) could also occur within the final product [29,30].…”

“…Additionally, unintentional or intentional use of cheaper (i.e., lower quality), less pure, or mislabeled raw materials can occur during the oil blending or extraction process [29,30]. Addition of unwanted additives or mislabeling (especially if a manufacturer produce several type of oil blends in parallel) could also occur within the final product [29,30]. This might also explain why some trace fatty acids (i.e, < 1 g/100 g) are lost when the standard mixture profile (n = 22 fatty acids) is compared with those profiles obtained in commercial samples (n = 7 fatty acids).…”

Transmittance Multispectral Imaging for Reheated Coconut Oil Differentiation

“…Therefore, it is of paramount importance to continuously monitor the quality of food. Rapid screening of food and beverages, including edible oils, has become a key focus among scientists and industrialists because contamination and adulteration of food compromise the quality of food [1]- [3]. In this context, the need for accurate, fast, non-destructive, and economical methods to assure the standard of food are of a timely need.…”