1H NMR chemometric models for classification of Czech wine type and variety

A set of 917 wines of Czech origin registered in a national competition were analysed using nuclear magnetic resonance spectroscopy (NMR) with the aim to build and evaluate multivariate statistical models and machine learning methods for the classification of type (5 types), variety (13 varieties) and location (4 locations) based on 1H NMR spectra. The predictive models afforded more than 93% of correctness for classification of dry and medium dry, medium, sweet white wines and dry red wines. The trained Random Forest (RF) method is able to classify Pinot noir with 96% of correctness, Blaufränkisch 96%, Riesling 92%, Cabernet Sauvignon 77%, Chardonnay 76%, Gewürtztraminer 60%, Hibernal 60%, Grüner Veltliner 52%, Pinot gris 48%, Sauvignon Blanc 45%, Pálava 40%. Pinot blanc and Chardonnay are varieties, which are often misplaced and were discriminated with 71% of correctness. Chemometrics represents prospective tool to predict important features in wine, potentially important in quality assessment and fraud detection.