[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: collinearity diagnostics for categorical and non-regression models
Kit Baum has uploaded a new version of -perturb- to SSC. Perturb adds
random noise to selected variables, then re-estimates the model.
Categorical variables are misclassified for a small percentage of
cases. This is repeated a number of times, after which the stability
of the estimates can be evaluated. Perturb thus forms an alternative
to collinearity diagnostics such as vif and condition indexes.
Besides being suitable for ordered and unordered variables, it can be
applied to models other than linear regression. Feedback welcome.
Kit also uploaded -coldiag2- to SSC. -coldiag2- is an improved
version of -coldiag- by Joseph Harkness. It uses -symeigen- of X'X
instead of -svd-, allowing it to work on much larger datasets. It
produces the same output as -coldiag-, condition indexes and variance
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
* For searches and help try: