20 Hyperparameters Quiz Questions and Answers

1. Question: What is a hyperparameter in machine learning?
Options:
A) A parameter that is learned during the training process.
B) A parameter that is set before the training process begins.
C) A parameter that represents the output of the model.
D) A parameter used only in unsupervised learning.
Answer: B
Explanation: Hyperparameters are configuration settings for a machine learning algorithm that are tuned prior to training, as they control the learning process and are not adjusted by the algorithm itself.

2. Question: Which of the following is an example of a hyperparameter in a neural network?
Options:
A) Weights of the connections.
B) Number of hidden layers.
C) The output predictions.
D) The activation values in the output layer.
Answer: B
Explanation: The number of hidden layers is set before training and influences the model’s architecture, making it a hyperparameter, unlike weights which are learned during training.

3. Question: What is the primary purpose of tuning hyperparameters?
Options:
A) To minimize the size of the dataset.
B) To optimize the model’s performance on unseen data.
C) To increase the number of features in the model.
D) To speed up the data preprocessing step.
Answer: B
Explanation: Tuning hyperparameters helps in achieving better generalization and reducing errors on new data by finding the optimal settings that balance bias and variance.

4. Question: In gradient descent, which hyperparameter controls how much the weights are updated in each iteration?
Options:
A) Batch size.
B) Learning rate.
C) Number of epochs.
D) Momentum.
Answer: B
Explanation: The learning rate determines the step size taken towards the minimum of the loss function, directly affecting the convergence speed and stability of the training process.

5. Question: Which technique is commonly used to search for the best hyperparameters?
Options:
A) Forward selection.
B) Grid search.
C) K-means clustering.
D) Linear regression.
Answer: B
Explanation: Grid search systematically explores a specified subset of hyperparameters by evaluating the model’s performance for each combination, helping to identify the optimal set.

6. Question: What does the batch size hyperparameter represent in training a model?
Options:
A) The total number of training examples.
B) The number of data samples processed before updating the model.
C) The size of the output layer.
D) The number of features in the input data.
Answer: B
Explanation: Batch size defines how many samples are used in one iteration of training, impacting memory usage, training speed, and the smoothness of the gradient descent updates.

7. Question: In decision trees, what is the role of the maximum depth hyperparameter?
Options:
A) It determines the minimum number of samples per leaf.
B) It limits the depth of the tree to prevent overfitting.
C) It sets the total number of trees in a forest.
D) It controls the learning rate of the tree.
Answer: B
Explanation: Maximum depth restricts how deeply the tree can be built, helping to control model complexity and reduce the risk of overfitting to the training data.

8. Question: Which hyperparameter is associated with regularization in linear models?
Options:
A) Alpha in Ridge regression.
B) Bias term.
C) Output threshold.
D) Feature scale.
Answer: A
Explanation: Alpha controls the strength of regularization, balancing the trade-off between fitting the data and keeping the model simple to avoid overfitting.

9. Question: How does increasing the number of epochs affect model training?
Options:
A) It reduces the model’s accuracy.
B) It allows the model more opportunities to learn from the data.
C) It decreases the batch size automatically.
D) It eliminates the need for hyperparameters.
Answer: B
Explanation: The number of epochs specifies how many times the entire dataset is passed through the model, potentially improving learning but risking overfitting if set too high.

10. Question: What is a potential downside of setting a very low learning rate?
Options:
A) The model trains too quickly and overfits.
B) The training process becomes unstable.
C) Convergence takes a long time or may not occur.
D) It increases the batch size.
Answer: C
Explanation: A low learning rate means smaller updates to the weights, which can lead to slower convergence and require more computational resources.

11. Question: In k-means clustering, what is the k hyperparameter?
Options:
A) The number of features to cluster.
B) The number of clusters to form.
C) The distance metric used.
D) The initialization method.
Answer: B
Explanation: K represents the number of clusters the algorithm will attempt to find, directly influencing the partitioning of the data.

12. Question: Which hyperparameter is used in dropout layers of neural networks?
Options:
A) Dropout rate.
B) Activation function.
C) Layer width.
D) Loss function.
Answer: A
Explanation: The dropout rate determines the probability of dropping out units during training, helping to prevent overfitting by promoting redundancy in the network.

13. Question: What effect does a high value of lambda in LASSO regression have?
Options:
A) It encourages more complex models.
B) It performs more feature selection by shrinking coefficients.
C) It speeds up the training process.
D) It reduces the number of epochs.
Answer: B
Explanation: Lambda controls the strength of L1 regularization, leading to sparser models by setting some coefficients to zero, thus performing feature selection.

14. Question: In support vector machines (SVM), what does the kernel hyperparameter define?
Options:
A) The learning rate.
B) The type of decision boundary.
C) The number of support vectors.
D) The batch size for training.
Answer: B
Explanation: The kernel defines how the input data is transformed into a higher-dimensional space, affecting the shape and separability of the decision boundary.

15. Question: Why is cross-validation important when tuning hyperparameters?
Options:
A) It eliminates the need for a test set.
B) It provides a more reliable estimate of model performance.
C) It reduces the model’s training time.
D) It automatically selects the best algorithm.
Answer: B
Explanation: Cross-validation assesses how the model performs on unseen data by splitting the dataset multiple times, helping to avoid overfitting during hyperparameter tuning.

16. Question: What is the momentum hyperparameter in optimizers like SGD?
Options:
A) It controls the initial learning rate.
B) It helps accelerate convergence by adding a fraction of the previous update.
C) It sets the number of layers in the network.
D) It determines the output activation.
Answer: B
Explanation: Momentum incorporates a portion of the previous gradient update, smoothing the path to the minimum and helping escape local minima.

17. Question: In random forest algorithms, what does the number of estimators hyperparameter refer to?
Options:
A) The depth of each tree.
B) The number of decision trees in the ensemble.
C) The size of the training data.
D) The regularization strength.
Answer: B
Explanation: The number of estimators specifies how many individual trees are built and combined, improving robustness and reducing variance.

18. Question: How does early stopping use hyperparameters?
Options:
A) By setting the patience parameter for monitoring validation loss.
B) By directly modifying the weights.
C) By changing the input data format.
D) By increasing the learning rate.
Answer: A
Explanation: The patience hyperparameter defines how many epochs to wait before stopping if the validation performance does not improve, preventing overfitting.

19. Question: Which hyperparameter might be tuned in a convolutional neural network (CNN)?
Options:
A) Filter size in convolutional layers.
B) The predicted class labels.
C) The input image pixels.
D) The final loss value.
Answer: A
Explanation: Filter size determines the receptive field in CNNs, affecting how features are extracted from the input data and influencing the model’s ability to detect patterns.

20. Question: What is the difference between hyperparameters and model parameters?
Options:
A) Hyperparameters are learned, while model parameters are set manually.
B) Model parameters are learned during training, while hyperparameters are set beforehand.
C) Both are the same and interchangeable.
D) Hyperparameters are only for unsupervised learning.
Answer: B
Explanation: Model parameters, like weights, are adjusted based on data during training, whereas hyperparameters are predefined and used to configure the learning algorithm.