Non-Destructive Geographical Traceability and Quality Control of Glycyrrhiza uralensis Using Near-Infrared Spectroscopy Combined with Support Vector Machine Model

TL;DR

This paper introduces a non-destructive method using near-infrared spectroscopy and machine learning to trace the geographical origin and assess the quality of licorice.

Contribution

A novel SVM-based framework for licorice traceability with over 99% accuracy using NIR spectroscopy.

Findings

The proposed method achieved over 99% accuracy in classifying licorice origins.

The framework is rapid, efficient, and environmentally friendly for quality control.

It provides a scientific basis for standardization in the functional food industry.

Abstract

Licorice (Glycyrrhiza uralensis Fisch.) is a widely used natural sweetener and functional food ingredient. Its sensory profile, nutritional value, and bioactive composition are strongly affected by geographical origin and cultivation mode, particularly the distinction between wild and cultivated resources. Consequently, developing a rapid and robust method for origin traceability is imperative for rigorous quality control and product standardization. This study proposes a non-destructive traceability framework integrating near-infrared (NIR) spectroscopy with a Support Vector Machine (SVM). The method’s validity was rigorously evaluated using a comprehensive dataset collected from China’s three primary production regions—Gansu Province, the Inner Mongolia Autonomous Region, and the Xinjiang Uygur Autonomous Region, encompassing both wild and cultivated resources. Experimental results demonstrated that the proposed framework achieved an overall classification accuracy exceeding 99%. The results show that the proposed method offers a rapid, efficient, and environmentally friendly analytical tool for the quality assessment of licorice, providing a scientific basis for rigorous quality control and standardization in the functional food industry.

Figures