The Kernel Density Estimate as a Measure of the Performance of One and Two-Dimensional TLC Systems with Large Retention Datasets in the Context of their Use in Fingerprinting

摘要

A new objective chromatographic response function, R-K, based on the kernel density estimate, is introduced for estimation of the fingerprinting performance of a particular TLC system (uniformity of retention) for which a large set of experimental R-F values of possible components of the mixture is available. The R-K criterion is insensitive to large numbers (hundreds or thousands) of R-F values, when the previously proposed criteria cease. It can be applied to one and two-dimensional TLC and is easily computed. As an example of its application, the performance of twelve general screening systems was evaluated in the context of herbal extract fingerprinting (88 phytochemical standards) by both one and two-dimensional TLC.