B. Grung et Om. Kvalheim, RESOLUTION OF MULTICOMPONENT PROFILES WITH PARTIAL SELECTIVITY - A COMPARISON OF DIRECT-METHODS, Chemometrics and intelligent laboratory systems, 29(1), 1995, pp. 75-87
Three procedures for resolution of multicomponent systems were tested
for their ability to deal with partial selectivity. The methods were e
volving factor analysis (EFA), heuristic evolving latent projections (
HELP) and the method of orthogonal projections. An extensive data simu
lation, in which five factors were varied according to a full 2(5) fac
torial design, was carried out to test the methods. The simulations sh
owed HELP to be the most powerful method for resolution. The importanc
e of the design variables was qualitatively studied by means of PLS mo
dels. Several parameters, among them the net analytical signal, were f
ound to be unable to alone say anything conclusively about HELP's chan
ces for successfully resolving the systems.