Model inputs are the predictor and response variables whose relationship is determined by a model fit. Input specifications supported by MachineShop are summarized in the table below.
formula | Traditional model formula |
matrix | Design matrix of predictors |
ModelFrame | Model frame |
ModelSpecification | Model specification |
recipe | Preprocessing recipe roles and steps |
Response variable types in the input specifications are defined by the user with the functions and recipe roles:
Response Functions | BinomialVariate |
DiscreteVariate | |
factor | |
matrix | |
NegBinomialVariate | |
numeric | |
ordered | |
PoissonVariate | |
Surv | |
Recipe Roles | role_binom |
role_surv |
Inputs may be combined, selected, or tuned with the following meta-input functions.
ModelSpecification | Model specification |
SelectedInput | Input selection from a candidate set |
TunedInput | Input tuning over a parameter grid |
fit
, resample