Thanks to Kit Baum, an update to the -pspline- package is available
from SSC. -pspline- provides a penalized spline scatterplot smoother
based on -xtmixed-.
To install the update, type
. ssc install pspline, replace
or use the -adoupdate- command.
Changes:
o To circumvent convergence problems in situations where there is
only little deviation in the data from a simple parametric model (e.g.
a linear model if degree=1, a quadratic model if degree=2), -pspline-
now performs a pilot goodness-of-fit (GOF) test for the parametric
model. The GOF test is implemented as a Wald test of the spline terms
in a non-penalized model. A low p-value indicates that there is a lot
of evidence against the parametric model. -pspline- now uses the
penalized spline model only if the p-value is smaller than 0.3 (or as
set by the -alpha()- option) and otherwise sticks with the parametric
model. A new -force- option allows you to skip the test and enforce
the penalized spline model.
o -pspline- has a new -discrete- option to fit a penalized model for
discrete variables with only few distinct values. If -discrete- is
specified, the variable is treated as a factor variable and a model
containing a random effect among the levels of the variable is used
instead of a spline. Example:
. sysuse auto
. pspline mpg rep78, discrete
o -pspline- has a new -at()- option to obtain the smooth at the
values of a specified variable (using linear interpolation).
o -pspline- now has a -replace- option to allow -generate()- overwite
existing variables.
ben
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
