Rapid evaluation of the comprehensive quality of Paeonol/Cyclodextrin supramolecular complexes using CASSA based on near-infrared spectroscopy combined with artificial intelligence algorithm

Preparing volatile component/cyclodextrin supramolecular complexes is a common method for enhancing the stability of volatile components. However, the quality assessment of supramolecular complexes is highly complex. This study first prepared paeonol/cyclodextrin supramolecular complexes and evaluated their overall quality. Then, Fourier Transform Near-Infrared Spectroscopy (FT-NIR) was combined with artificial intelligence (AI) to build a Support Vector Machine Classification (SVM) model and a Partial Least Squares Regression (PLSR) model. The SVM classification model reached 100% accuracy, whereas the PLSR quantitative model demonstrated R² > 0.90 for both calibration and prediction sets. Results confirm that integrating FT-NIR with AI improves the accuracy and reliability of qualitative/quantitative models. Via the Comprehensive Analysis of Single Spectral Acquisition (CASSA) method, dual detection of the complexes’ formation state and concentration was achieved, enabling rapid, comprehensive quality evaluation. This study demonstrates the excellent prospects of combining FT-NIR technology with the preparation of supramolecular complexes.