Optimization of Detection Wavelelengths in Liquid Chromatography on the Basis of Information Theory and Kaiman Filter

Abstract

An optimization method of detection wavelengths in column chromatography is described on the basis of the information measure FUMI (FUnction of Mutual Information) which denotes the mutual information of chromatography. Functions required as quality criteria for success in the wavelength optimization comprise the total mutual information di (FUMI), the information loss 8i and the information variation of individual peaks AJi. The variation AJi is defined by slight modification of FUMI and is shown to be indispensable for the wavelength optimization. An experimenatal design for column chromatographic determination of antipyretics mixtures, previously proposed, is evaluated information-theoretically in terms of the three functions.