Geographical origin differentiation of Philippine Robusta coffee (C. canephora) using X-ray fluorescence-based elemental profiling with chemometrics and machine learning

Abstract

The increasing demand for authenticity and traceability in high-value crops underscores the need for reliable methods to verify the geographical origin of single-origin coffee and prevent fraud. This study explores a rapid and cost-effective approach utilizing Energy-Dispersive X-ray Fluorescence (EDXRF) elemental profiling combined with chemometrics and machine learning techniques. The concentrations of ten elements (K, P, Ca, S, Cl, Fe, Cu, Mn, Sr, Zn) were analyzed in 43 green Robusta coffee samples from four Philippine provinces to assess origin differentiation. Principal Component Analysis (PCA) revealed distinct clustering patterns, while Linear Discriminant Analysis (LDA) achieved 79 % classification accuracy. Random Forest (RF) improved accuracy to 84 %, highlighting its potential for geographical classification. This study serves as a proof of concept for employing XRF-based elemental profiling to differentiate Robusta coffee by origin, providing baseline data to support the development of authenticity and traceability systems within the Philippine coffee industry.