A lab dataset has 10,000 samples, 200 features, and 5 labeled classes. You want to visualize the data in 2D before classification. Which approach is most appropriate?

Question

Answer

K-Means with K=2 to reduce to 2 cluster assignments.

Answer

EM/GMM with 2 components to project data probabilistically.

Answer

LDA to 2D, because it leverages class labels to find the most discriminative projection, making class separability visible in the 2D plot.

Answer

PCA to 2D, since it is always preferable to LDA for high-dimensional data.

Crowdly

A lab dataset has 10,000 samples, 200 features, and 5 labeled classes. You want ...

Want instant access to all verified answers on moodle.spit.ac.in?