Laser induced breakdown spectroscopy for elemental analysis and discrimination of honey samples

In the present work Laser Induced Breakdown Spectroscopy (LIBS) is employed for the classification of honey samples assisted by different machine learning algorithms. Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Support Vector Machines (SVMs) and Random Forest Classifiers (RFCs) were used for the treatment of the LIBS spectroscopic data, while the advantages and the suitability of each statistical analysis technique is discussed. It was found that the spectral lines of the main inorganic constituents of honey: Ca, Mg, Na and K are the most important for classification purposes. In all cases, excellent classification results were obtained, attaining remarkable accuracies exceeding 95%. The present results suggest the potential use of the LIBS technique assisted by machine learning algorithms for honey classification based on its floral origin, providing an easy to use and efficient methodology able to perform real time quality control. [Display omitted] •Laser-Induced Breakdown Spectroscopy (LIBS) for honey analysis.•Honey classification based on botanical origin by LIBS assisted by machine learning.•PCA, LDA, SVM, RFC models for honey classification/discrimination.•Classification accuracies exceeded 95% for all models.•RFC reveals Mg, Na, Ca and K lines' importances for classification.