This study (open access) used a novel 2-stage chemometric approach to classify Extra Virgin Olive Oil (EVOO) vs Virgin Olive Oil (VOO) based upon an untargeted database of Visible and Near Infra-Red (NIR) spectra. Differentiating EVOO from VOO is one of the more difficult tasks of a sensory panel, and the authors propose that their new process is used as a screen before sensory testing.
In the first step, a one-class model is used to test whether the sample belongs to the target class of (EVOO + VOO) or whether it does not (i.e., that it is Lampeter Olive Oil, LOO). This is based on a PLS-DM model.
In the second step, a discriminant model was applied to those samples identified as “target class” to test if they better fitted as EVOO or VOO.
The authors conclude that their proposed approach is uniquely suited to addressing the analytical question “Differentiate between EVOO and VOO”.
Comments