Origin discrimination of protected geographical indication (PGI) Sunite lamb using fingerprinting technology coupled...

This study employed stable isotope technology (δ¹³C, δ¹⁵N, δ²H, δ¹⁸O), mineral element technology (Na, Mg, P, Cl, K, Ca, Cr, Fe, Cu, Zn, As, Se, Sr, Mo, Cd, Sb, Pb), and their combined technology to establish an origin discrimination model for Protected Geographical Indication (PGI) Sunite lamb. A total of 211 samples were collected according to different levels of geographical distance from PGI Sunite lamb. Four machine learning algorithms-Linear Discriminant Analysis (LDA), Random Forest (RF), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN)-were used to develop the model. The identification accuracy ranked as: combined technology (94.12%) > stable isotope technology (92.94%) > mineral element technology (89.41%). For algorithms, performance followed: RF and KNN > SVM > LDA. The optimal model integrating the combined technology with the RF algorithm achieved an overall accuracy of 94.12%, exhibiting excellent performance for the origin discrimination of PGI Sunite lamb.