Skip to main content
Evidence-Based Supplement Research
Evidence-Based Supplement Research

Study Design

Methods
Developed a multimodal data fusion analytical framework integrating Raman spectroscopy and low-field NMR, with variable selection methods (CARS, SPA, UVE) and correlation analyses (Pearson, Spearman, Kendall). Multiple chemometrics were employed to discriminate ten types of oil samples and predict olive oil content in mixtures.
Funding
Unclear
Enabling rapid and accurate detection of high-quality olive oil for combating adulteration, this study developed a multimodal data fusion analytical framework integrating Raman spectroscopy and low-field nuclear magnetic resonance (LF-NMR). Key features derived from both datasets were extracted using variable selection methods, including competitive adaptive reweighted sampling (CARS), successive projections algorithm (SPA), and uninformative variable elimination (UVE), and correlation analyses, including Pearson, Spearman and Kendall. Based on distinct inputs, multiple chemometrics were employed to discriminate ten types of oil samples and predict olive oil content in the mixtures. The extreme gradient boosting (XGBoost) model achieved 90.74% classification accuracy, utilizing Raman UVE and LF-NMR SPA, and yielded R2 of 0.942 for binary and 0.909 for ternary blends, utilizing CARS-selected features from both techniques. SHapley Additive exPlanations (SHAP) analysis was applied to reveal the decisive spectral and relaxation features. The data fusion strategy provides a robust, high-precision framework for olive oil authentication.

Research Insights

    Back to top