Optimization of gradient separation for chromatographic fingerprint of herbal medicine: a predictive model combined with grid search.

A four-step strategy is described which is aimed at optimizing the separation parameters of the chromatographic fingerprint of herbal medicines. In the optimization procedure, a model is built to predict the retention time of components. The proposed model shows good retention prediction potency with an average relative deviation of 1.78%. The grid search method is used in optimizing gradient parameters. In addition, an iterative procedure is adopted to further improve the prediction accuracy of the gradient retention time and the quality of the chromatogram. The whole optimization process has been validated by real samples.