Laser-induced breakdown spectroscopy assisted chemometric methods for rice geographic origin classification.

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Chemometrics

Background:

• Food adulteration and mislabeling pose significant challenges in the food industry.
• Laser-induced breakdown spectroscopy (LIBS) offers advantages for material analysis due to its speed and minimal sample preparation requirements.
• Chemometric methods are essential for extracting meaningful information from complex spectral data.

Purpose of the Study:

• To evaluate the performance of various chemometric algorithms for classifying rice based on geographic origin using LIBS data.
• To identify the most accurate and efficient method for rapid rice origin classification.
• To assess the potential of LIBS-chemometric approaches for detecting food adulteration.

Main Methods:

• Laser-induced breakdown spectroscopy (LIBS) was used to analyze 20 types of rice samples from different geographic origins without pretreatment.
• Principal Component Analysis (PCA) was applied for dimensionality reduction and collinearity management of spectral data.
• Several chemometric algorithms, including Decision Tree (DT), Random Forest (RF), Partial Least Squares Discriminant Analysis (PLS-DA), Linear Discriminant Analysis (LDA), and Support Vector Machine (SVM), were employed for classification.

Main Results:

• LDA achieved the highest classification accuracy (98.60% with 89 variables, 98.40% with 30 principal components) and the fastest operation time (2.09 s with 89 variables, 0.36 s with 30 principal components).
• Support Vector Machine (SVM) also demonstrated high accuracy (99.20% with 89 variables, 99.20% with 30 principal components) but with significantly longer operation times.
• Cross-validation results further supported LDA's efficiency and accuracy, with 98.35% accuracy using 30 principal components.

Conclusions:

• Linear Discriminant Analysis (LDA) is the most effective and efficient chemometric tool for classifying rice geographic origins when coupled with LIBS.
• The LIBS-LDA approach provides a rapid, accurate, and reagent-free method for identifying adulterated agricultural products.
• This technique holds significant potential for ensuring food authenticity and safety in the agricultural sector.