Answers are grounded in this paper and cite their sources.
Developing global classification and calibration models for identifying edible oil type and predicting peroxide value, a measure of the degree of oxidation directly related to oil freshness is predicated on capturing variance differences within and between growing seasons across a large set of 19 naturally aged (5–7 years) oil types. Our approach utilizes a larger sample set and longer aging period than reported studies (typically 2–5 years), includes multiple growing seasons, and is based on natural rather than accelerated thermal aging of oil samples. We provide spectroscopic data (Raman scattering, near- and mid-infrared absorption) for 19 edible oil types, which represent the most diverse and only naturally aged sample set reported to date. The data, which consist of three data collection campaigns acquired using different analytical methods at different times during the aging process, provide opportunities to investigate and validate new machine learning approaches for classification, calibration, and data fusion. In the supplemental material, vibrational spectroscopy data from all samples are included along with the R programming language code for preprocessing data and building basic classification and calibration models.
No takes yet. Share an insight, caveat, or question.
Gilbraith et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: