Data Analysis & Modeling Techniques Flashcards

In association rule mining, which measure indicates how much more often items A and B occur together than expected if they were independent?

Answer: Lift

Lift = confidence(A→B) / support(B); values above 1 indicate positive association.

A model scores 98% training accuracy but 71% on held-out data. What is the most likely issue?

Answer: Overfitting

A large gap between training and validation performance is the classic sign of overfitting.

Which clustering algorithm can identify arbitrarily shaped clusters and flag noise points without requiring the number of clusters in advance?

Answer: DBSCAN

DBSCAN groups density-connected points and labels low-density points as noise.

Which impurity measure is used by the CART algorithm for classification splits by default?

Answer: Gini index

CART uses the Gini index for classification trees and variance reduction for regression trees.

What is the main purpose of principal component analysis (PCA)?

Answer: Reduce dimensionality while retaining maximal variance

PCA projects data onto orthogonal components that capture the greatest variance.

In k-fold cross-validation with k=10, how many times is each observation used for validation?

Answer: Once

Each observation appears in exactly one validation fold and in k−1 training folds.

Which statement about multicollinearity in linear regression is correct?

Answer: It inflates the variance of coefficient estimates

Highly correlated predictors make coefficient estimates unstable with large standard errors.