Machine Learning for Robotics - Decision Tree

What is Decision Tree in the context of Machine Learning?

• A Decision Tree is a type of neural network used for image recognition tasks.

• A Decision Tree is a clustering algorithm that groups data points based on similarity.

• A Decision Tree is a reinforcement learning technique that uses rewards to make decisions.

• A Decision Tree is a machine learning algorithm used for classification and regression tasks by recursively splitting the data based on features to create a tree-like structure of decisions.

Explain the concept of entropy in Decision Trees.

• Entropy in Decision Trees is a measure of impurity or disorder in a dataset. It is used to decide the best split at each node by calculating the entropy before and after the split. The goal is to minimize entropy and maximize information gain.

• Entropy is calculated by summing the squared differences between predicted and actual values

• Entropy is only used in regression models, not decision trees

• Entropy is a measure of the number of features in a dataset

What are the advantages of using Decision Trees in Robotics?

• High computational complexity

• Requires extensive data preprocessing

• Interpretability, handling non-linear relationships, less data preprocessing, handling numerical and categorical data

• Limited interpretability

How does a Decision Tree handle missing values in data?

• By randomly imputing missing values with the mode of the feature

• By excluding data points with missing values from the model

• By assigning a default value to all missing values

• By finding the best split based on available features and using existing splits to guide predictions for data points with missing values.

Discuss the process of pruning in Decision Trees.

• Pruning in Decision Trees refers to trimming the leaves of the tree to change its appearance.

• Pruning involves watering the decision tree regularly to help it grow faster.

• Pruning is the process of adding more branches to the decision tree to increase accuracy.

• Pruning involves either pre-pruning, where the tree is pruned as it is being built, or post-pruning, where the tree is built first and then pruned. Common pruning techniques include reduced error pruning, cost complexity pruning, and pessimistic error pruning.

What is the difference between Gini impurity and Entropy in Decision Trees?

• Gini impurity always leads to overfitting, while Entropy does not.

• Gini impurity is more suitable for regression problems compared to Entropy.

• Entropy is faster than Gini impurity in decision tree calculations.

• Gini impurity is computationally faster, while Entropy tends to produce more balanced trees.

Explain the concept of Information Gain in Decision Trees.

• Information Gain is calculated by taking the entropy of the child nodes minus the entropy of the parent node.

• Information Gain is not used in Decision Trees.

• Information Gain is calculated by taking the entropy of the parent node minus the weighted average of the entropy of the child nodes after the split.

• Information Gain is the same as Gini Impurity in Decision Trees.

How can overfitting be prevented in Decision Trees?

• Adding more features to the model

• Reducing the training data size

• Ignoring missing values

• Pruning the tree, setting minimum samples to split, setting maximum depth, or using ensembling techniques like Random Forest.

What are some common algorithms used to build Decision Trees?

• Random Forest

• K-Means

• ID3, C4.5, CART, Gini impurity

• Entropy

Discuss a real-world application of Decision Trees in Robotics.

• Speech recognition, obstacle avoidance, and sensor fusion

• Object recognition, path planning, and autonomous navigation