Classification of persimmon fruit origin by near infrared spectrometry and least squares-support vector machines

• Single spectrometric analysis for origin recognition of fruit samples.
• Evaluating several chemometric approaches to find the best appropriate one.
• Proposing a rapid reliable approach which could be maintained for industrial applications.

The main objective of this work has been the authentication by Fourier transform near infrared (FT-NIR) spectrometry of the origin of persimmon fruits cultivated in different regions of Spain. In order to achieve this goal, 166 persimmon samples from 7 different regions of Spain were analyzed by FT-NIR spectrometry. By splitting the spectral data in training and independent test sets, a classification model was built using least squares support vector machines chemometric technique. Orthogonal signal correction and principal component analysis were performed prior to conduct the classification strategy. The verified model was applied for the prediction of the origin of 50 samples from the independent test set being the prediction error 2%.