initial support for GLRM imputation
prauc_table
auc_table - default null hypothesis is now the highest AUC.
csvl_auc_table() - table of AUCs for each learner, plus Discrete SL and SL.
factors_to_indicators
allow missing values in factors
h2o_init_multinode - create a multi-node h2o cluster e.g. on a SLURM system.
impute_missing_values
support integer64 numeric type (data.table)
impute_missing_values
support supplying imputation values from another dataset, for use when scoring a model to a larger dataset.
load_all_code() - changed default environment from baseenv() to .GlobalEnv, which will make explicit package references unnecessary.
missingness_indicators - return matrix of integers rather than numerics.
set_java_memory() - specify maximum memory usage for rJava JVM.
get_java_memory() - get maximum memory allocated to a rJava JVM.
Mode - option to not choose NA as a mode, set TRUE by default.
plot_roc - plot the best learner, not necessarily the SuperLearner.
sl_h2o_auto - automatic machine learning via h2o.
sl_mgcv - add mgcv wrapper for splines.
sl_xgboost_cv - integrate support for gaussian family outcomes.
vim_corr - new function to rank covariates by their (possibly weighted) univariate correlation with the outcome.
don't drop matrix to a vector, e.g. when using analyzing a single covariate.
impute_missing_values
fix missingness indicators when no indicators need to be created.
missingness_indicators
Fix when only one column of missingness indicators needs to be created.
plot_roc - fix axis labels for updated ggplot2.
Initial release on CRAN.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.