MS-DS Master of Data science Master of Data Science 1

Question 1

Which ensemble learning method trains multiple models sequentially, where each model attempts to correct the errors made by the previous one?

Accepted Answer

Boosting

Answer

Bagging

Answer

Random Forest

Answer

Stacking

Question 2

In Principal Component Analysis (PCA), what do the eigenvalues of the covariance matrix represent?

Accepted Answer

The amount of variance explained by each principal component

Answer

The direction of maximum variance in the data

Answer

The correlation between original features

Answer

The mean of each transformed feature

Question 3

What distinguishes stochastic gradient descent (SGD) from batch gradient descent?

Accepted Answer

SGD updates model parameters using one (or a few) training examples per iteration rather than the entire dataset

Answer

SGD uses a higher learning rate by default

Answer

SGD computes the exact gradient of the loss function

Answer

SGD is only applicable to convex loss functions

Question 4

In the context of a Master of Data Science program, what does 'data leakage' during model training refer to?

Accepted Answer

When information from the test set inadvertently influences the training process

Answer

Unauthorized access to sensitive training datasets

Answer

Loss of training data due to storage failures

Answer

Overfitting caused by too many features

Question 5

Which probability theorem is foundational to generative classifiers like Naive Bayes, allowing posterior class probabilities to be computed from likelihoods and priors?

Accepted Answer

Bayes' Theorem

Answer

Central Limit Theorem

Answer

Law of Large Numbers

Answer

Jensen's Inequality

Question 6

What is the vanishing gradient problem in deep neural networks, and which architecture feature was specifically introduced to mitigate it?

Accepted Answer

Gradients shrink exponentially as they propagate to earlier layers; mitigated by residual (skip) connections

Answer

Gradients become too large during backpropagation; mitigated by dropout layers

Answer

Weights converge to zero due to L2 regularization; mitigated by batch normalization

Answer

The loss surface becomes flat near the optimum; mitigated by adaptive learning rates