Multivariate statistical methods (chemometrics, taxometrics) are ideal for evaluating chemotaxonomic data. Some of these methods are best used for unsupervised classification approaches such as cluster analysis, multidimensional scaling, correspondence analysis, and PCA, whereas other methods such as Partial Least Squares Discriminant (PLS-D) analysis and soft independent modeling of class analogy (SIMCA) are more suited for discriminant analysis and identification (Frisvad 1994b; Soderstrom and Frisvad 1984).
The immense quantities of data collected by modern analytic instruments dictates some form of automatic data handling and analysis (Nielsen et al. 1999) although problems such as handling of simultaneous UV, MS, and nuclear magnetic resonance data without component identification needs to be solved, as well as issues relating to the storage in searchable databases, and how data can be combined with other biodiversity information.
Was this article helpful?