Some recent features/changes:
- The license was changed to GPL >= 2 to accommodate new code from the
- New feature selection functions
safswere added, along with helper functions and objects, were added. The package HTML was updated to expand more about feature selection. I’ll talk more about these functions in an upcoming blog post.
- A reworked version of
nearZerVarbased on code from Michael Benesty was added the old version is now called
nzvthat uses less memory and can be used in parallel.
sbfControlnow has a
multivariateoption where all the predictors are exposed to the scoring function at once.
- Several regression simulation functions were added:
- For the input data
train, we now respect the class of the input value to accommodate other data types (such as sparse matrices).
- A function
Recently added models:
- From the
adabagpackage, two new models were added:
- Weighted subspace random forests from the
wsrfpackage was added.
- Additional bagged FDA and MARS models were added (model codes
bagEarthGCV) were added that use the GCV statistic to prune the model. This leads to memory reductions during training.
- Brenton Kenkel added ordered logistic or probit regression to
method = "polr"from
- The adaptive mixture discriminant model from the
- A robust mixture discriminant model from the
robustDApackage was added.
- The multi-class discriminant model using binary predictors in the
bindapackage was added.
- Ensembles of partial least squares models (via the
enplspackage) was added.
- From the
kernlabpackage, SVM models using string kernels were added:
- The model code for
adahad a bug fix applied and the code was adapted to use the “sub-model trick” so it should train faster.