1. Which of the following is true about the "learning rate" in gradient descent?
A. A higher learning rate may lead to faster convergence but could overshoot the
optimal solution.
B. A lower learning rate will result in a model that converges too quickly.
C. The learning rate has no effect on the model's training.
D. The learning rate should always be set to 1 for optimal performance.
Answer: a) A higher learning rate may lead to faster convergence but could overshoot
the optimal solution.
Rationale: A learning rate that's too high may cause the model to skip over the optimal
solution, while a smaller rate ensures gradual convergence but can be slower.
2. What is the purpose of the "dropout" technique in deep learning?
A. To increase the size of the training dataset
B. To reduce overfitting by randomly setting some of the model’s weights to zero
during training
C. To speed up the model training process
D. To normalize the output of the model
,Answer: b) To reduce overfitting by randomly setting some of the model’s weights to
zero during training
Rationale: Dropout helps prevent overfitting by randomly disabling some neurons
during training, forcing the model to learn more robust features.
3. Which of the following is a disadvantage of using decision trees?
A. They are highly interpretable and easy to visualize.
B. They are prone to overfitting, especially with complex trees.
C. They require large amounts of data to perform well.
D. They are not capable of handling numerical data.
Answer: b) They are prone to overfitting, especially with complex trees.
Rationale: Decision trees can easily overfit the training data, especially when the tree
is allowed to grow too deep or complex without pruning.
4. What is the "elbow method" used for in machine learning?
A. To find the optimal number of clusters in a clustering algorithm
B. To determine the best value for the learning rate
C. To evaluate the model's performance
D. To decide the best regularization parameter
, Answer: a) To find the optimal number of clusters in a clustering algorithm
Rationale: The elbow method helps determine the ideal number of clusters in
algorithms like K-means by identifying the point where adding more clusters doesn't
significantly improve the model.
5. Which of the following techniques can be used for dimensionality reduction?
A. K-nearest neighbors
B. Principal Component Analysis (PCA)
C. Decision Trees
D. Logistic Regression
Answer: b) Principal Component Analysis (PCA)
Rationale: PCA is a method for reducing the number of features in a dataset while
preserving as much variance as possible, making it an effective technique for
dimensionality reduction.
6. Which of the following techniques can be used for feature selection?
A. Cross-validation
B. Recursive Feature Elimination (RFE)
C. K-means clustering
A. A higher learning rate may lead to faster convergence but could overshoot the
optimal solution.
B. A lower learning rate will result in a model that converges too quickly.
C. The learning rate has no effect on the model's training.
D. The learning rate should always be set to 1 for optimal performance.
Answer: a) A higher learning rate may lead to faster convergence but could overshoot
the optimal solution.
Rationale: A learning rate that's too high may cause the model to skip over the optimal
solution, while a smaller rate ensures gradual convergence but can be slower.
2. What is the purpose of the "dropout" technique in deep learning?
A. To increase the size of the training dataset
B. To reduce overfitting by randomly setting some of the model’s weights to zero
during training
C. To speed up the model training process
D. To normalize the output of the model
,Answer: b) To reduce overfitting by randomly setting some of the model’s weights to
zero during training
Rationale: Dropout helps prevent overfitting by randomly disabling some neurons
during training, forcing the model to learn more robust features.
3. Which of the following is a disadvantage of using decision trees?
A. They are highly interpretable and easy to visualize.
B. They are prone to overfitting, especially with complex trees.
C. They require large amounts of data to perform well.
D. They are not capable of handling numerical data.
Answer: b) They are prone to overfitting, especially with complex trees.
Rationale: Decision trees can easily overfit the training data, especially when the tree
is allowed to grow too deep or complex without pruning.
4. What is the "elbow method" used for in machine learning?
A. To find the optimal number of clusters in a clustering algorithm
B. To determine the best value for the learning rate
C. To evaluate the model's performance
D. To decide the best regularization parameter
, Answer: a) To find the optimal number of clusters in a clustering algorithm
Rationale: The elbow method helps determine the ideal number of clusters in
algorithms like K-means by identifying the point where adding more clusters doesn't
significantly improve the model.
5. Which of the following techniques can be used for dimensionality reduction?
A. K-nearest neighbors
B. Principal Component Analysis (PCA)
C. Decision Trees
D. Logistic Regression
Answer: b) Principal Component Analysis (PCA)
Rationale: PCA is a method for reducing the number of features in a dataset while
preserving as much variance as possible, making it an effective technique for
dimensionality reduction.
6. Which of the following techniques can be used for feature selection?
A. Cross-validation
B. Recursive Feature Elimination (RFE)
C. K-means clustering
