The supported models currently all come from tidypredict right now.
The following models are supported by tidypredict:
lm()glm()randomForest::randomForest()ranger -
ranger::ranger()earth::earth()xgboost::xgb.Booster.complete()Cubist::cubist()partykit -
partykit::ctree()parsniptidypredict supports models fitted via the
parsnip interface. The ones confirmed currently work in
tidypredict are:
lm() - parsnip: linear_reg()
with “lm” as the engine.randomForest::randomForest() - parsnip:
rand_forest() with “randomForest” as the
engine.ranger::ranger() - parsnip:
rand_forest() with “ranger” as the engine.earth::earth() - parsnip:
mars() with “earth” as the engine.The following 46 recipes steps are supported
step_BoxCox()step_adasyn()step_bin2factor()step_bsmote()step_center()step_corr()step_discretize()step_downsample()step_dummy()step_filter_missing()step_impute_mean()step_impute_median()step_impute_mode()step_indicate_na()step_intercept()step_inverse()step_lag()step_lencode_bayes()step_lencode_glm()step_lencode_mixed()step_lincomb()step_log()step_mutate()step_nearmiss()step_normalize()step_novel()step_nzv()step_other()step_pca()step_pca_sparse()step_pca_sparse_bayes()step_pca_truncated()step_range()step_ratio()step_rename()step_rm()step_rose()step_scale()step_select()step_smote()step_smotenc()step_sqrt()step_tomek()step_unknown()step_upsample()step_zv()