Machine Learning Concepts Quiz

<p>When the dataset is imbalanced, with a significantly larger number of instances for one class. (A)</p> Signup and view all the answers

<p>Confusion Matrix (D)</p> Signup and view all the answers

<p>The test set is specifically designed to be representative of real-world data, making its performance a better indicator of the model's effectiveness in practical applications. (A)</p> Signup and view all the answers

<p>A binary classification problem is a specific type of problem useful for prediction, often involving two possible outcomes. (D)</p> Signup and view all the answers

<p>To predict a continuous value, such as price or quantity. (B), To improve the model's performance on unseen data. (D), To identify patterns and relationships within the data. (F)</p> Signup and view all the answers

<p>A lower MSE generally indicates a better model fit. (B)</p> Signup and view all the answers

<p>Regression models use a continuous target variable, while classification models use a discrete target variable. (C)</p> Signup and view all the answers

<p>To find a function that accurately predicts the output for new inputs, even those not in the training set. (B)</p> Signup and view all the answers

<p>To validate the performance of the learned model on unseen data and assess its ability to generalize. (B)</p> Signup and view all the answers

<p>It represents the inherent uncertainty or variability in the relationship between inputs and outputs, which cannot be perfectly captured by the model. (C)</p> Signup and view all the answers

<p>The use of a pre-existing model or algorithm to estimate the relationship between inputs and outputs. (A)</p> Signup and view all the answers

<p>It measures the model's ability to perform well on new data, beyond the training set, making it applicable to real-world scenarios. (D)</p> Signup and view all the answers

<p>The estimated function learned by the algorithm, which predicts the output 𝑦ො based on the input values. (B)</p> Signup and view all the answers

<p>They are used to identify patterns and trends in the input-output relationship, which the algorithm then uses to construct a function that maps inputs to outputs. (C)</p> Signup and view all the answers

<p>It represents the specific features or attributes that are used to characterize the data examples. (A)</p> Signup and view all the answers

<p>It provides a more accurate estimate of the model's true performance in real-world scenarios where previously unseen data is encountered. (D)</p> Signup and view all the answers

<p>To ensure that the model can predict outputs for inputs that are not included in the training set, making it useful in practical applications. (D)</p> Signup and view all the answers

<p>True Positive (C)</p> Signup and view all the answers

<p>Accuracy would be low, as no positive examples would be predicted correctly. (A)</p> Signup and view all the answers

<p>It defines the threshold for predicting a class label. (A)</p> Signup and view all the answers

<p>Accuracy = (TP + TN) / Total Observations (B)</p> Signup and view all the answers

<p>Fitting the model with training data (B)</p> Signup and view all the answers

<p>All of the above (D)</p> Signup and view all the answers

<p>The number of wins against a human opponent. (C)</p> Signup and view all the answers

<p>Machine Learning is a subset of Artificial Intelligence. (B)</p> Signup and view all the answers

<p>A calculator that can perform complex mathematical calculations. (A)</p> Signup and view all the answers

<p>The learning algorithm is guided by a set of predefined rules. (A)</p> Signup and view all the answers

<p>A set of predetermined rules that guide the learning process. (D)</p> Signup and view all the answers

<p>Support Vector Machines (SVMs) used for classification problems. (A)</p> Signup and view all the answers

<p>It requires the machine to physically resemble a human. (C)</p> Signup and view all the answers

<p>Reinforcement learning (B)</p> Signup and view all the answers

<p>The unknown function that relates input variables to the output. (A)</p> Signup and view all the answers

<p>Development of new algorithms and tools (C)</p> Signup and view all the answers

<p>Supervised learning uses labeled data, while unsupervised learning uses unlabeled data. (A)</p> Signup and view all the answers

<p>Training a model to predict house prices based on features like size and location. (D)</p> Signup and view all the answers

<p>The agent learns from labeled data. (C)</p> Signup and view all the answers

<p>To find hidden patterns and relationships in data. (A)</p> Signup and view all the answers

<p>Predicting the price of a stock tomorrow. (C)</p> Signup and view all the answers

<p>It allows for the processing of larger datasets and complex algorithms. (A)</p> Signup and view all the answers

<p>IBM Watson's win in Jeopardy (A)</p> Signup and view all the answers

A type of machine learning where models learn from labeled data to make predictions.

A procedure that identifies patterns in data and creates a model for predictions.

A learned model that distinguishes different classes within data.

Field allowing computers to learn without explicit programming, identifying patterns from data.

Machine learning programs enhance performance on tasks by learning from past experiences.

An evaluation to determine if a machine can exhibit intelligent behavior indistinguishable from a human.

A specific challenge or objective that a machine learning model is designed to learn from.

A method or metric used to evaluate how well a machine learning model performs on a task.

Landmark achievements in AI include Checkers and Jeopardy by IBM Watson, and Deep Blue vs. Garry Kasparov in Chess.

There are three main categories: Supervised, Unsupervised, and Reinforcement Learning.

A machine learning category using labeled data and direct feedback to predict outcomes.

A machine learning type that operates without labeled data, focusing on discovering hidden structures.

A machine learning approach that uses a reward system to learn a series of actions in an environment.

Predicting continuous outcomes based on input variables in supervised learning.

Categorizing data into discrete classes based on input variables in supervised learning.

Data that includes input-output pairs, used in supervised learning to train models.

In unsupervised learning, there are no explicit targets or feedback during training.

In reinforcement learning, an agent interacts with the environment based on the state, chooses an action, and receives rewards.

The actual outcome measured from training data, denoted as 𝑦.

The outcome estimated by the model using input variables, denoted as 𝑦ො.

The estimated function that maps input variables to predicted outputs.

Pairs of input-output used to train a learning algorithm.

The ability of a model to perform well on unseen test examples.

The features or independent variables used in predicting the outcome.

The dependent variable or the result that is predicted by the model.

Data used to evaluate the model's predictive performance after training.

The noise or discrepancies in data that affect predictions.

A table used to evaluate the performance of a classification model by showing true and false predictions.

The count of positive examples correctly predicted as positive by the model.

The count of positive examples incorrectly predicted as negative by the model.

A metric that measures the ratio of correct predictions to the total observations in a model.

A graphical representation of a classifier's performance across different thresholds, illustrating the trade-off between true positive rate and false positive rate.

A predictive problem that predicts categorical outcomes, often binary.

A predictive problem that focuses on predicting continuous outcomes.

A metric to evaluate regression models by calculating average squared errors.

The balance between a model's simplicity (bias) and its ability to fit data noise (variance).

When a model is too simple and fails to capture relationships in the data.

When a model is too complex and fits noise in the training data instead of generalizing.

The proportion of correctly predicted observations out of total observations.

The test set is used to evaluate the model's performance on unseen data, as opposed to the training set used for building the model.

A dataset where classes are not represented equally, which can mislead accuracy metrics.