Excitation-emission matrix fluorescence spectroscopy coupled with multi-way chemometric techniques for characterization and classification of Chinese lager beers

•Fluorescence EEMs and chemometrics were used to classify Chinese lager beers.•Multiple feature extraction and classification techniques were evaluated.•PARAFAC was used to decompose and characterize the multi-way data arrays.•A novel four-way data was constructed to solve the classification problem.•PARAFAC-data fusion-kNN and four-way PARAFAC-kNN methods obtained satisfactory results. This paper proposed excitation-emission matrix fluorescence spectroscopy coupled with multi-way chemometric techniques for characterization and classification of Chinese pale lager beers produced by different manufacturers. The undiluted and diluted beer samples presented different fluorescence fingerprints. Three-way and four-way parallel factor analysis (PARAFAC) were used to decompose the skillfully constructed three-way and four-way data arrays, respectively, to further achieve beer characterization and feature extraction. Based on the features extracted in different ways, four strategies for beer classification were proposed. In each strategy, three supervised classification methods including linear discriminant analysis (LDA), partial least squares discriminant analysis (PLS-DA) and k-nearest neighbor (kNN) were used to build discriminant models. By comparison, PARAFAC-data fusion-kNN method in strategy 3 and four-way PARAFAC-kNN method in strategy 4 obtained the best classification results. The classification strategy based on four-way sample-excitation-emission-dilution level data array was proposed to solve the problem of beer classification for the first time.