Quiz: Fundamentals of Machine Learning — 9 domande

1. What is the primary purpose of a machine learning model?

A. To visualize data patterns

B. To replace traditional programming entirely

C. To create predictive functions based on data

D. To store data efficiently

Answer: (C)

Spiegazione

The main purpose of a machine learning model is to create predictive functions that can estimate unknown outcomes or values based on past data. It learns from data to make predictions or classifications.

2. Which algorithm is commonly used to perform simple linear regression in machine learning?

A. Gradient descent

B. Logistic regression

C. Ordinary Least Squares

D. K-Means clustering

Answer: (C)

Spiegazione

Ordinary Least Squares is the classic algorithm used to fit linear regression models by minimizing the sum of squared residuals. Gradient descent can also be used but is a more general optimization method.

3. Which of the following best describes the process of training a machine learning model?

A. Using the model to predict new data without prior learning

B. Manually programming rules for data classification

C. Collecting data without any processing

D. Applying an algorithm to find relationships between features and labels in data

Answer: (D)

Spiegazione

Training involves applying an algorithm to the dataset to find the relationship between features (x) and labels (y), thereby deriving a function y = f(x) that can be used for predictions.

4. What is the primary difference between supervised and unsupervised learning?

A. Supervised learns from labeled data; unsupervised finds patterns in unlabeled data

B. Supervised only predicts categories; unsupervised predicts continuous values

C. Supervised uses neural networks; unsupervised uses decision trees

D. Supervised is faster than unsupervised; unsupervised is more accurate

Answer: (A)

Spiegazione

Supervised learning requires labeled data to train models for prediction, while unsupervised learning finds structures and patterns within unlabeled data.

5. In the context of machine learning, what distinguishes deep learning from traditional models?

A. Deep learning is only used for image processing

B. Deep learning models do not require training data

C. Deep learning models are simpler and less accurate

D. Deep learning uses neural networks with multiple layers and mimics biological neurons

Answer: (D)

Spiegazione

Deep learning employs neural networks with layered architectures that mimic biological neurons, allowing for complex pattern recognition and learning from large datasets, unlike traditional shallow models.

6. Which evaluation metric is most appropriate for assessing a classification model's performance in distinguishing between categories?

A. Mean Absolute Error (MAE)

B. Accuracy

C. Mean Squared Error (MSE)

D. R-squared (R2)

Answer: (B)

Spiegazione

Accuracy measures the proportion of correct predictions and is commonly used to evaluate classification models, unlike MAE, MSE, or R2, which are more suited for regression tasks.

7. In deep learning, neural networks typically use what kind of architecture to extract complex patterns?

A. Single-layer perceptron

B. Layered architectures with multiple hidden layers

C. K-Nearest Neighbors

D. Decision trees

Answer: (B)

Spiegazione

Deep learning employs neural networks with multiple hidden layers, allowing models to learn complex representations of data.

8. According to the revision sheet, which of the following is NOT a common machine learning algorithm?

A. Linear regression

B. K-Means clustering

C. Decision forests

D. Logistic regression

Answer: (C)

Spiegazione

Decision forests (or random forests) are indeed common algorithms, but the term 'Decision forests' was not listed explicitly; thus, it's a trick question. Given the sheet, all listed are common; 'Decision forests' refers to an ensemble of decision trees, so it's actually correct. However, since the options are meant to test knowledge, the safest answer is 'Decision forests' as the least explicitly mentioned with this exact name in the sheet.

9. What is the main goal during the training phase of a machine learning model?

A. To collect new data

B. To evaluate the model's performance

C. To adjust model parameters to minimize loss

D. To make predictions on unseen data

Answer: (C)

Spiegazione

Training involves adjusting model parameters, such as weights in a neural network, to minimize the loss function using the training data, which enables the model to make accurate predictions.