The Representation of Chemical Spectral Data for Classification

摘要

The classification of unknown samples is among the most common problems found in chemometrics. For this purpose, a proper representation of the data is very important. Nowadays, chemical spectral data are analyzed as vectors of discretized data where the variables have not connection, and other aspects of their functional nature e.g. shape differences (structural), are also ignored. In this paper, we study some advanced representations for chemical spectral datasets, and for that we make a comparison of the classification results of 4 datasets by using their traditional representation and two other: Functional Data Analysis and Dissimilarity Representation. These approaches allow taking into account the information that is missing in the traditional representation, thus better classification results can be achieved. Some suggestions are made about the more suitable dissimilarity measures to use for chemical spectral data.