Machine Learning Concepts Quiz

A science where computers learn without explicit programming, improving with experience.

Machine learning with labeled data; the machine learns under guidance like a student with a teacher.

Machine learning with unlabeled data; the machine finds patterns and structures without a teacher.

The dataset used to train the machine learning model.

Data with associated categories or labels.

Data without categories or labels.

A supervised learning algorithm that classifies data based on nearest neighbors.

A supervised learning algorithm that uses a tree-like model to make decisions from data.

KNN is useful for classification problems with manageable features (less than 20) and plenty of training data.

KNN has no training period, is easy to implement, and new data integrates smoothly.

KNN slows down with many queries and is vulnerable to misleading features.

Decision trees start with a root node and branch downward in a tree structure, splitting data based on attributes until reaching a leaf node with a prediction (class label).

Decision Tree training involves choosing relevant attributes, creating a table of past experiences showing target attributes and building a decision tree via an algorithm, like ID3, to progressively sort data into classes.

A type of machine learning where an agent learns by interacting with an environment, taking actions, and receiving rewards or penalties.

A machine learning method where the algorithm is trained on labelled data, mapping inputs to known outputs.

A machine learning method used to identify patterns and structure in unlabeled data without pre-defined outputs.

Data with known output values or labels, used to train supervised learning algorithms.

Data without pre-defined output values or labels used for unsupervised learning.

A learning method where an agent tries different actions and learns from the consequences (rewards or penalties)

A supervised learning problem type where the algorithm classifies inputs to predefined classes.

An unsupervised learning problem type where data is grouped into clusters based on similarities.

A supervised learning problem type where the algorithm predicts a continuous output value.

A subfield of unsupervised learning that finds relationships or rules between variables in databases.

A supervised learning algorithm that uses a tree-like structure to make decisions based on data.

A node in a decision tree that tests an attribute (feature) of the data.

A characteristic or feature of the data being analyzed.

A node in a decision tree that holds a decision or class (e.g., YES/NO).

In the example, weather condition (e.g., sunny, overcast, rain).

The amount of water vapor in the air used for decision making.

The decision variable in the example (play tennis or not).

Machine learning with labeled data, where the algorithm learns from provided examples.

A supervised learning method that uses a tree-like structure to classify data based on attributes.

The attribute chosen to create different branches in the decision tree, separating the data based on a certain condition.

A function used to evaluate the quality of a splitting attribute in a decision tree.

A goodness function used in decision tree algorithms to select the best attribute to split on, based on reducing uncertainty about the target variable.

Building a decision tree by recursively partitioning the training data based on attributes.

Removing parts of a decision tree to increase accuracy on unseen data.

Input data used to evaluate the decision tree's prediction on new occurrences.

Process of assessing attributes to determine the best one for splitting data in a decision tree, typically done using a goodness function.

Difference in entropy before and after a split in a dataset.

Measure of uncertainty in a dataset.

Portion of dataset where a specific attribute has a particular value.

Average considering different values for each dataset subset by their size.

Used to calculate entropy (-probability*log2(probability)).

Categorical attribute in the decision tree dataset.

Information gain from using Outlook to split the dataset.

Information gain from using Temperature to split the dataset.

Information gain from using Humidity to split the dataset.

Information gain from using Windy to split the dataset.

Starting point of the decision tree.

A tree-like model for making decisions based on data.

Final part of the decision tree; prediction made here.

The rules used to divide dataset subsets.