Abstract:
ABSTRACT
Laser-induced autofluorescence (LIAF) was used to characterize and classify
some selected commercial anti-malarial herbal drugs (AMHDs) to avoid sample
destruction. Results from deconvoluted peaks of the LIAF spectra of AMHDs
showed secondary metabolites belonging to derivatives of alkaloids and classes of phenolic compounds (tannins, lignin, flavonoid, and coumarins) present in all the AMHDs samples. Analyses with unsupervised methods (Principal component
analysis (PCA), K-means clustering and Hierarchical Clustering Analysis (HCA))
classified the AMHDs that were similar in physicochemical properties. Based on
two PCs, supervised pattern recognition methods such as Supervised Vector
Machine (SVM), K-Nearest Neighbor (KNN), Quadratic Discriminant Analysis
(QDA), and Linear Discriminant Analysis (LDA) models generated were used to
identify and classify unknown AMHDs with 100.00, 100.00, 99.52 and 99.04 %
accuracy respectively. LIAF technique in combination with multivariate analysis
may offer non-destructive characterization and classification of AMHDs.
