MLA_Lec2

Which of the following accurately characterizes a decision tree and its components?

• • A network-based model where nodes are interconnected and data flows through multiple paths.

• A tree-like model where internal nodes represent decisions based on attribute tests, branches represent possible outcomes, and leaf nodes represent final classifications or predictions.

• • A hierarchical structure where internal nodes represent attribute values, branches represent data splits, and leaf nodes represent predictions.

• • A linear model where data is sequentially processed through nodes.

Entropy is a measure of:

• • The accuracy of a classification model

• The impurity or randomness of a dataset

• The certainty of a dataset

• The speed of a machine learning algorithm

A decision tree created with the aim to achieve very minimal entropy at each node is likely to suffer from:

• High variance
  

• Overfitting

• Underfitting

• High bias

The goal of feature selection while building a decision tree is to find features or attributes (decision nodes) which lead to split in children nodes which has:

• • No preference for entropy

• • Maximum entropy

• • Equal entropy

• Minimum entropy

Which of the following is NOT a component of the bias-variance decomposition?

• Overfitting

• Bias

• Variance

• Bayes error

Information gain is a measure of the ______ in information obtained about the target variable through the splitting of data based on an attribute.

The decision tree algorithm typically selects the attribute with the ______ information gain as the splitting criterion at each node.

The depth of a node in a decision tree is the length of the ______ from the root node to that node.

What is the depth of the tree below? What is the depth of node X4 in the tree below?

• - The depth of the tree is :

• - The depth of the node X4 in the tree below is :

Pruning a decision tree is always a necessary step to prevent overfitting and improve generalization performance

• False

• True