Two-Way Data Analysis: Multivariate Curve Resolution – Iterative Resolution Methods

摘要: This chapter describes the general modus operandi of model-free multivariate curve resolution (MCR) iterative methods, that is, recovery pure concentration profiles and responses (spectra) from optimization initial estimates under action constraints. Initially, some common concepts, such as methodologies for generation or description most constraints, are presented. The basic bilinear (CR) model is expressed in two different forms: X = CST + E X = (CoR)(R−1SoT) + E. Methods based on first equation solve C and/or ST matrices directly, whereas methods second optimize transformation matrix R a way C = CoR ST = R−1SoT chemically meaningful. Resolving factor analysis (RFA) through elementary transformations (Gentle) selected algorithms representative via described detail this chapter. direct, solution target (ITTFA) resolution-alternating least squares (MCR-ALS). Critical comments graphical examples use applicability also included.