Electronic Resource
New York, NY
:
Wiley-Blackwell
Journal of Chemometrics
9 (1995), S. 211-221
ISSN:
0886-9383
Keywords:
diagnostics statistics
;
QSAR
;
MASCA
;
principal component regression
;
non-least squares regression
;
types of multicollincarity
;
flagged observations
;
influential points
;
high-leverage points
;
outliers
;
extra-carrier points
;
random perturbation
;
cluster correalation
;
resampling
;
Chemistry
;
Analytical Chemistry and Spectroscopy
Source:
Wiley InterScience Backfile Collection 1832-2000
Topics:
Chemistry and Pharmacology
Notes:
The formal application of a Hansch analysis to a series of 3-quinuclidinyl benzylates (QNBs) led to a ‘statistically significant’ QSAR equation. In contrast, the application of the MASCA model has shown that the design matrix is unsuitable for each QSAR analysis: one sample member is an outlier but not a high-leverage or influential point; another one is an influential point, a high-leverage point and an extra-carrier point. The regressors of the design matrix are multicollinear without predictive model power. The result of such flagged observation and this type of multicollinearity is a multiple cluster correlation. The QNB series is a good example for ‘sampling artifacts’ where no practically important but artificial QSARs can be found.
Additional Material:
3 Ill.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1002/cem.1180090307
Permalink
|
Location |
Call Number |
Expected |
Availability |