ML2 GPT2 Medium

What is the primary assumption of LDA?

• a) Normal distribution of predictors

• b) Identical covariance matrices

• c) Presence of outliers

• d) Linear separation

What is the role of eigenvalues in PCA?

• a) Determine class labels

• b) Indicate the amount of variance captured by each principal component

• c) Maximize between-class variance

• d) Minimize within-class variance

What is the primary function of Standard Deviation in descriptive statistics?

• a) Average value

• b) Spread of data

• c) Mode

• d) Median

What is the primary limitation of PCA?

• a) Cannot handle non-linear data

• b) Sensitivity to outliers

• c) Requires class labels

• d) Ignores variance

What is the primary purpose of Quadratic Discriminant Analysis (QDA)?

• a) Minimize within-class variance

• b) Assume identical covariance matrices

• c) Adapt better to different covariance structures for each class

• d) Ignore class labels

In SVM, how does a large value of C affect the decision boundary?

• a) Smaller margin

• b) Larger margin

• c) Decrease overfitting

• d) No impact

When would you prefer to use Empirical Probability?

• a) Always

• b) When observations are not available

• c) When known quantities and assumed distributions are available

• d) Based on observations or experiments

When is PCA considered less useful?

• a) When variables are correlated

• b) When data is non-linear

• c) When variance needs to be captured

• d) When variables are uncorrelated as its primary function is to capture variance explained by correlations among variables

What does the Interquartile Range (IQR) measure in descriptive statistics?

• a) Average distance from the mean

• b) Spread of data

• c) Variance

• d) Measure of variability robust to extreme outliers