Differentiation of Pueraria lobata and Pueraria thomsonii using partial least square discriminant analysis (PLS-DA)

- Partial least square discriminant analysis models were established.
- Variables were selected from the entire chromatogram and puerarin.
- Genetic algorithm and successive projection algorithm provide comparable classification ability to the entire chromatogram.
- The entire chromatogram provides superior predictability in determining raw material used in commercial granules.
- The entire chromatogram is more adequate than puerarin as the quality control criteria in the authentication of P. lobata.