Instructions

For this homework, you should create a new folder in your homework directory. Call it HW5 or something similar that you can keep track of. Download the homework markdown template file Student_HW_template.Rmd from the course webpage, and put a copy in this folder. Rename it something like HW5_YourName.Rmd. This markdown document will be where you will answer each of the questions below.

The Assignment

We’re going to deal again with the titanic data from your in class activity.

titanic <- read_csv("titanic.csv") %>% 
  clean_names() %>%
  mutate(survived = as.factor(survived))

EDA

You did this already in your activity, but I’m leaving this section here because it’s an important step in the process: Always vizualize before modeling!

Modeling

1. Perform a train/test split and create a cross-validation split. Build a recipe to preprocess this data to use in a logistic model and KNN model.

2. Specify a Logistic Regression model for classification, using the glmnet engine. Set the penalty and mixture arguments to tune().

3. Specify a KNN model for classification, using the kknn engine. Set the neighbors and dist_power arguments to tune().

4. Create a workflow set containing your recipe(s) and two model specifications. You will need to look up in the TMWR book how to do that (Section 11.1, and Chapter 15 will be the most help).

Tuning

1. Using the workflow_map function, run the tune_grid function on your workflow set and choose the best parameters for each model.

2. Use collect_metrics to look at the accuracy estimates for each model. Which family of model (KNN or LogReg) is performing better, if either?

3. Finalize your model by choosing the best performing one. Fit the model to the training set and report your test set accuracy.

Submitting HW

When you’ve successfully answered all the questions, knit your document to a PDF file. Look through it to make sure everything worked the way you expect it to. You will submit both your .Rmd and .pdf files to Schoology.