A small shiny app to explore a simple decision tree model (rpart) on titanic training data.
Only three predictor variables are displayed: gender, age, and passenger class.
Interactive options:
Choose whether surrogate variables are used to substitute missing predictor values (here, more variables are being used in the background as surrogates).
Select how the first split is determined:
By the algorithm (which selects gender);
Enforce age;
Enforce passenger class.