The generalized rank annihilation method (GRAM) has been criticised for not
having a global least squares fitting property such as the alternating lea
st squares (ALS) method. In Pan 1 of this series, we have modified GRAM by
introducing a weight for the data matrices. The proposed modification is ca
lled iteratively reweighted GRAM (IRGRAM). Here, it is shown that these wei
ghts enable one to shed new light on the least squares fitting properties o
f GRAM and ALS. Inequalities are derived which suggest that IRGRAM compares
favourably with ALS in terms of model fit to the data matrices. Although a
pplying different weights directly affects the sums of squares explained by
IRGRAM and ALS, error propagation shows that the first-order approximation
to prediction variance remains unaltered when using IRGRAM. In contrast, t
he effect on the variance in the estimated profiles depends on the analyte
under consideration. This result suggests that the amount of Fitted data do
es not give a clear indication of the performance of bilinear calibration m
odels. (C) 2001 Elsevier Science B.V. All rights reserved.