Exhaustive evolving factor analysis (E‐EFA)

Abstract

Abstract Reliable information about concentration windows is crucial for many self‐modeling curve resolution (SMCR) methods. Algorithms based on evolving factor analysis (EFA)‐type methods enable the detection of the beginnings and endings of concentration profiles. Often this is sufficient to define concentration windows. However, under certain circumstances the connection of beginnings and endings is ambiguous, and wrong connection results in erroneous analysis. Systematic rank analysis of all possible submatrices of a matrix of data can result in unambiguous identification of the correct connectivity of beginnings and endings of concentration profiles. Steps are proposed which accelerate the calculations dramatically. Copyright © 2001 John Wiley & Sons, Ltd.