Rapid detection of thiabendazole in food using SERS coupled with flower-like AgNPs and PSL-based variable selection algorithms

Thiabendazole (TBZ) exposure through food can have substantial long-term health consequences for humans. Herein, SERS coupled with flower-like AgNPs and python-scikit-learn (PSL) algorithms, was proposed for rapid detection of TBZ. Initially, AgNFs with a strong enhancement factor (EF) of 1.303 × 106 were synthesized to collect TBZ spectra. Subsequently, three PSL-based variable selection algorithms, including SelectKBest (SKB), variance threshold (VT) and recursive feature elimination (RFE), were comparatively applied to select the informative variables for TBZ prediction. The RFE exhibited the optimal selection ability, and RFE-SVM achieved the best performance for TBZ prediction (Rp2 = 0.976, RPD = 6.477), with a computed limit of detection (LOD) of 0.24 μg/mL obtained. Finally, good recoveries of spiked samples (78.5–95.09%) were obtained, demonstrating that the proposed method is practicable and potentially effective for TBZ rapidly detection in food.