Decision tree learning

What is the main principle behind decision tree learning?

• Memorizing training examples

• Building a flowchart of decisions

• Maximizing overfitting

• Predicting randomly

In decision tree representation, each internal node corresponds to:

• A class label

• A feature or attribute

• A training sample

• A final decision

Which of the following is an appropriate problem for decision tree learning?

• Classifying spam emails

• Predicting future stock prices using time series

• Image recognition with raw pixel values

• Solving linear algebra equations

The decision tree algorithm uses which of the following strategy to grow the tree?

• Bottom-up construction

• Top-down greedy approach

• Randomized leaf selection

• Neural backpropagation

What does the “hypothesis space” in decision tree learning represent?

• The number of classes in the dataset

• All possible tree structures

• The training data points

• The maximum depth of a tree

What is the inductive bias of a decision tree learner?

• Preference for deeper trees

• Preference for trees that split on numerical features

• Preference for shorter trees with higher information gain

• No bias at all

Which of the following is a common issue with decision tree learning?

• Too little memory usage

• High accuracy with small datasets

• Overfitting to training data

• Cannot handle categorical features

Which metric is commonly used to select the best feature to split in decision trees?

• Learning rate

• Information Gain

• Gradient error

• K-means clustering score

What does pruning a decision tree help prevent?

• Underfitting

• Infinite recursion

• Overfitting

• High training time

In a basic Python implementation using scikit-learn, which class is used to build a decision tree?

• sklearn.neural.DecisionNet

• sklearn.tree.DecisionTreeClassifier

• sklearn.ensemble.RandomTree

• sklearn.linear_model.TreeClassifier