Quality assessment of olive oils based on temperature-ramped HS-GC-IMS and sensory evaluation: Comparison of different processing approaches by LDA, kNN, and SVM.

For the first time, this study describes a HS-GC-IMS strategy for analyzing non-targeted volatile organic compounds (VOCs) profiles to distinguish between virgin olive oils of different classification. Correlations among measured flavor characteristics and sensory attributes evaluated by a test panel were determined by applying unsupervised (PCA, HCA) and supervised (LDA, kNN and SVM) chemometric techniques. PCA and HCA were applied for natural clustering of the samples and LDA, kNN, and SVM methods were used to create predictive models for olive oil classification. Identification of 26 target compounds revealed which compounds are responsible for discrimination, and how their distribution correlates with the sensory evaluation. In the investigated samples, LDA, kNN, and SVM models correctly classified 83.3%, 73.8%, and 88.1% of the samples, respectively. This suggests that mathematical correlations of HS-GC-IMS 3D fingerprints with the sensory analysis may be appropriate for calculating a good predictive value to classify virgin olive oils.