Advancing from unsupervised, single variable-based to supervised, multivariate-based methods: A challenge for qualitative analysis

This article reviews and describes the open challenges of defining the unreliability limit or region when advancing from unsupervised single variable-based to supervised, multivariate-based methods applied for the purpose of qualitative analysis. An unambiguous definition of unreliability regions is difficult to make when dealing with multivariate methods, although useful additional information, such as increased selectivity, may be gained when applying such methods.