Imagine a world where medicine isn't just about trust but also about proof. Traditional Chinese medicine, especially Atractylodis Rhizoma (AR), is celebrated for its healing properties. However, purity is always a concern. Enter hyperspectral imaging, a tool that can swiftly spot adulterations in AR powder. This study combines this tech with chemometric methods to create models that can detect and measure adulteration with high accuracy.

The key player here is Partial Least Squares Discriminant Analysis (PLS-DA), which builds classification models that are nearly perfect, boasting accuracy rates above 99%. For quantifying adulteration levels, methods like Partial Least Squares Regression (PLSR), Random Forest Regression (RFR), and BP Neural Network (BPNN) come into play. Among these, BPNN stands out for its stability.

What makes these models impressive? They all scored high on statistical measures like R-square (over 0. 97) and low on errors like Root Mean Square Error (RMSE, less than 0. 0300). Even after narrowing down to key wavelengths using algorithms like IRIV, SPA, and VISSA, the models maintained their high accuracy and reliability.

These findings underscore the potential of hyperspectral imaging and chemometrics in safeguarding the integrity of traditional medicine, ensuring every dose delivers what it promises.

Unveiling Hidden Truths: How Spectral Technology Saves Traditional Medicine

Actions