New stuff in
ftp://ftp.stat.ucla.edu/pub/xlispstat/code
in
/mathematics/optimization/annealing.lsp
Simulated Annealing following Corana et al, and Goffe
et al. (Jan de Leeuw)
/mathematics/optimization/examples/rosenbrock.lsp
Rosenbrock test-functions for optimization. (Jan de Leeuw)
/programming/make-proto/construct-methods.lsp
Automagically writes :isnew and assesssor methods for
a given prototype and a given list of slots. (Jan de Leeuw)
/programming/utilities/edit/spreadsheet.lsp
Spreadsheet data editor (upgrade).(Jan de Leeuw)
/statistics/simulation/model-selection.lsp
Jackknife based model selection for multinomial and
multinormal methods (in progress). (Jan de Leeuw)
/statistics/introstat/show-data
Module to interface data and plot methods. (Jason Bond)
/statistics/smoothers/gamfit.lsp
Fits generalized additive models. (Jason Bond)
/statistics/smoothers/super-smoother.lsp
Jerome Friedman's supersmoother (upgrade). (Jason Bond)
A continuing project is to port Common Lisp to Xlisp. This is usually
trivial for numerical and statistical stuff, as long as it is not
object oriented and does not rely on sophisticated macro functions
and other far-out parts of CLtL-2.
A new project, which shows up in
/mathematics/nswc
/mathematics/numrecipes
/mathematics/lsqpack
is to port numerical libraries from FORTRAN to CL, using the f2cl
translator from the University of Waikato. Then we tweak it a bit
and make it Xlisp. The stuff that is currently in the directories
I mentioned has not been tried out really, and it is probably
buggy. Use at your own risk, but please report bugs. Waikato will
start work on f2cl again, and they would surely like to hear about
bugs.
--- Jan
--
Jan de Leeuw; UCLA Statistics Program; UCLA Statistical Consulting
US mail: 8118 Math Sciences, 405 Hilgard Ave, Los Angeles, CA 90024-1554
phone (310)-825-9550; fax (310)-206-5658; email: ·······@stat.ucla.edu