Project Two Directions

The goal of this project is to create several different classification models and select the model that best predicts the response variable in a future data set the model has never seen.

1. Form a group of no fewer than two and no greater than four classmates.
2. Select one of the following data sets from the modeldata package:

Data	Predict
credit_data	Status
attrition	Attrition
leaf_id_flavia	species
mlc_churn	churn
taxi	tip

3. Create at least three classification models using different specifications from the parsnip package.
4. Create train and test sets.
5. Use ten-fold cross-validation to tune all hyperparameters in your models.
6. Report appropriate metrics (AUC, accuracy, etc.) for your trained models.
7. Use the provided Quarto template as a start to creating a 6 to 10 minute presentation. Hide code you do not want to show during your presentation; but, make sure all the code you used to create the models is embedded in the slides.

• Your presentation should include:

  1. The models you evaluated and their performance metrics on the test set.
  2. The steps you performed in your recipe.
  3. An evaluation of the time versus benefit of each model’s creation.