model-nnet.md
In tidypredict: Run Predictions Inside the Database

returns the right output

Code
  rlang::expr_text(tf)
Output
  [1] "0.7 + case_when(0.1 + wt * 0.2 + hp * 0.3 < -15 ~ 0, 0.1 + wt * \n    0.2 + hp * 0.3 > 15 ~ 1, .default = 1/(1 + exp(-(0.1 + wt * \n    0.2 + hp * 0.3)))) * 0.8 + case_when(0.4 + wt * 0.5 + hp * \n    0.6 < -15 ~ 0, 0.4 + wt * 0.5 + hp * 0.6 > 15 ~ 1, .default = 1/(1 + \n    exp(-(0.4 + wt * 0.5 + hp * 0.6)))) * 0.9"

an ordered factor is rejected

Code
  tidypredict_fit(model)
Condition
  Error in `acceptable_lm()`:
  ! The treatment contrast is the only one supported at this time. Field(s) with an invalid contrast are: "f".

multiple non classification outputs are rejected

Code
  tidypredict_fit(model)
Condition
  Error in `parse_model()`:
  ! `tidypredict_fit()` does not support `nnet::nnet()` models with multiple outputs that are not a classification.

matrix interface fits are rejected

Code
  tidypredict_fit(model)
Condition
  Error in `parse_model()`:
  ! `tidypredict_fit()` does not support `nnet::nnet()` models fit with the matrix interface.
  i Refit the model with the formula interface, so that the names of the predictors are available.

Code
  parse_model(model)
Condition
  Error in `parse_model()`:
  ! `tidypredict_fit()` does not support `nnet::nnet()` models fit with the matrix interface.
  i Refit the model with the formula interface, so that the names of the predictors are available.

tidypredict_test errors for classification nnet models

Code
  tidypredict_test(model, iris)
Condition
  Error in `tidypredict_test()`:
  ! `tidypredict_test()` does not support classification `nnet::nnet()` models.
  i Use `tidypredict_fit()` directly for class predictions.

inline functions in the formula are rejected

Code
  tidypredict_fit(model)
Condition
  Error in `acceptable_lm()`:
  x Functions inside the formula are not supported.
  i Functions detected: "log".  Use `dplyr` transformations to prepare the data.