gslnls 1.4.2
confint(), summary() and other methods no
longer fail in case of a singular gradient- Fixed bug: missing confidence/prediction intervals in
predict() for multi-start gsl_nls() call with
fn defined as a function in combination with
newdata.
gslnls 1.4.1
- Fixed compatibility GSL versions < 2.5
gslnls 1.4.0
- Robust loss optimization added in
gsl_nls() via
argument loss
weights in gsl_nls() accepts a matrix (in
addition to a vector) in which case the objective function is
generalized least squares- Added new function
gsl_nls_loss()
- Added new method
cooks.distance()
- Minor changes in
predict() and hatvalues()
for weighted NLS
gslnls 1.3.3
- Fix standard errors
predict() when using
newdata
gslnls 1.3.2
- Reverted to static Makevars.win (supplied by T. Kalibera)
- Added new method
hatvalues()
gslnls 1.3.1
- Minor edits configure.ac to fix cran check results
gslnls 1.3.0
- Missing starting values/ranges allowed in
gsl_nls()
lower and upper parameter constraints
included in gsl_nls()- Added 3 regression problems from Bates & Watts (1988)
- Updated multi-start algorithm in
gsl_nls()
- Added configure.win, cleanup.win and Makevars.win.in
- Removed old Makevars and Makevars.win
- Several minor changes
gslnls 1.2.0
- Added multi-start algorithm to
gsl_nls()
- Added 56 NLS regression and optimization test problems
- Added unit tests in folder
unit_tests
- Several minor changes/fixes
gslnls 1.1.1
- Clean exits
gsl_nls() and gsl_nls_large()
when interrupted
- Default algorithm in
gsl_nls_large() set to
"lm"
gslnls 1.1.0
- Added large-scale NLS regression with
gsl_nls_large()
gslnls 1.0.2
gslnls 1.0.1