parsnip contains wrappers for a number of models. For example, the parsnip function rand_forest() can be used to create a random forest model. The mode of a model is related to its goal. Examples would be regression and classification.

The list of models accessible via parsnip is:

classification: boost_tree(), decision_tree(), logistic_reg(), mars(), mlp(), multinom_reg(), nearest_neighbor(), null_model(), rand_forest(), svm_poly(), svm_rbf()

regression: boost_tree(), decision_tree(), linear_reg(), mars(), mlp(), nearest_neighbor(), null_model(), rand_forest(), surv_reg(), svm_poly(), svm_rbf()

How the model is created is related to the engine. In many cases, this is an R modeling package. In others, it may be a connection to an external system (such as Spark or Tensorflow). This table lists the engines for each model type along with the type of prediction that it can make (see predict.model_fit()).

model engine class classprob conf_int numeric pred_int quantile raw
boost_tree() C5.0 × × × ×
spark × × × ×
xgboost × × ×
decision_tree() C5.0 × × × ×
rpart × × ×
spark × × × ×
linear_reg() glmnet × × × × ×
keras × × × × × ×
lm × × ×
spark × × × × × ×
stan × × ×
logistic_reg() glm × × ×
glmnet × × × ×
keras × × × × ×
spark × × × × ×
stan × ×
mars() earth × × ×
mlp() keras × × ×
nnet × × ×
multinom_reg() glmnet × × × ×
keras × × × × ×
spark × × × × ×
nearest_neighbor() kknn × × ×
null_model() parsnip × × ×
rand_forest() randomForest × × ×
ranger × ×
spark × × × ×
surv_reg() flexsurv × × × × ×
survreg × × × × ×
svm_poly() kernlab × × ×
svm_rbf() kernlab × × ×

Models can be added by the user too. See the “Making a parsnip model from scratch” vignette.