Branch Out: A Beginner’s Quiz on Decision Trees, Splitting, and Gini Index Quiz

1. Understanding Split Criteria

   Which metric is commonly used to determine the best split in a decision tree when performing binary classification?

   1. A. Gini Index
   2. B. Mean Squared Error
   3. C. Support Vector
   4. D. Sigmoid Loss
   5. E. Fourier Transform

2. Basics of Splitting

   When constructing a decision tree, which action is taken to grow the tree after evaluating a node's data?

   1. A. Increasing regularization
   2. B. Splitting the node based on a selected feature
   3. C. Averaging node values
   4. D. Pruning the entire branch
   5. E. Removing the node immediately

3. Gini Index Values

   What is the Gini Index of a node containing only one class, such as all 'Yes' outcomes?

   1. A. 0.0
   2. B. 0.25
   3. C. 0.5
   4. D. 1.0
   5. E. 2.0

4. Pruning Purpose

   Why is pruning applied to a decision tree after it has grown to its full depth?

   1. A. To add more branches
   2. B. To improve overfitting by reducing complexity
   3. C. To remove the root node
   4. D. To ensure all features are used
   5. E. To maximize training data size

5. Splitting Example

   Given a dataset where splitting on 'Color' leads to two groups: all red objects in one and all blue objects in another, which property does this demonstrate?

   1. A. Class purity after splitting
   2. B. Random forest creation
   3. C. Min-max scaling
   4. D. Feature normalization
   5. E. Data balancing error

6. Gini Calculation Basics

   If a split results in two nodes, each containing half 'Yes' and half 'No' values, what does this say about the Gini Index of these nodes?

   1. A. The Gini Index is at its maximum
   2. B. The Gini Index is zero
   3. C. The Gini Index is negative
   4. D. The Gini Index does not change
   5. E. The Gini Index becomes infinite

7. Overfitting in Trees

   What is the risk of growing a decision tree without any restrictions on depth or minimum samples at nodes?

   1. A. Overfitting to noise in the training data
   2. B. Model always underfits
   3. C. Missing variables
   4. D. Increasing bias error
   5. E. Improving generalization

8. Leaf Nodes

   In the context of decision trees, what typically characterizes a leaf node?

   1. A. It performs further splits
   2. B. It contains the final class prediction
   3. C. It stores all features
   4. D. It averages input data
   5. E. It triggers pruning automatically

9. Choosing Features for Splitting

   When deciding which feature to split on at each step in building a decision tree, what is usually maximized or minimized?

   1. A. The number of input features
   2. B. The Gini gain or impurity decrease
   3. C. The absolute feature magnitude
   4. D. The feature's name alphabetically
   5. E. The feature's p-value

10. Pruning Types

    Which of the following is an example of post-pruning in decision trees?

    1. A. Removing branches after the tree is fully grown
    2. B. Avoiding splits during tree construction
    3. C. Selecting features at random before building the tree
    4. D. Normalizing data before splitting
    5. E. Adding more splits after training