A non-destructive classification method of peaberry and non-peaberry coffee using near infrared spectroscopy and machine learning algorithms

Coffee is a globally significant commodity, with increasing demand for specialty varieties such as peaberry coffee. Peaberry beans have distinct morphology and chemical composition, making them highly valued. However, manual separation is labor-intensive and inconsistent, necessitating automated classification methods. This study evaluated Near Infrared (NIR) Spectroscopy combined with machine learning algorithms, including Support Vector Machine (SVM), Random Forest (RF), and Linear Discriminant Analysis (LDA), for peaberry coffee beans classification. Spectral data were preprocessed using second derivative transformation and feature selection with the Boruta algorithm. Results showed SVM achieved 100 accuracy, while RF improved from 88.89 to 100 after feature selection, and LDA reached 97.92. These findings confirmed that the NIR spectra using machine learning algorithms provide a rapid, non-destructive, and highly accurate classification method. This approach might enhanced sorting efficiency, ensured product consistency, and increased the commercial value of peaberry coffee, contributing to a more competitive and sustainable coffee industry. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2026.