Reduction of ion mobility spectrometry data by clustering characteristic peak structures

Abstract

Abstract Spectrometric methods generate a large amount of data that are difficult to process due to high dimensionality and considerable variability. Therefore, a method for the detection of characteristic peak structures was established on the base of clustering techniques and a merging regions algorithm that enabled the reduction of data in the size of one million to about one hundred meaningful values with respect to fluctuating axis allocations. The effectiveness of the method was exemplified considering a set of ion mobility spectrometry measurements of human breath. In questions of comparative nature, the reduction even could be upgraded by multiple test procedures. Afterwards statistical analyses could be arranged easily for various targets. Copyright © 2007 John Wiley & Sons, Ltd.