K-Means clustering is described as what kind of model?

• A. Clustering
• B. Regression
• C. Binary
• D. Classification

Answer: (D)

The main idea being tested is how K-Means fits into the types of learning models: it’s an unsupervised method that forms groupings based on similarity, rather than predicting labels or values.

K-Means works by partitioning data into a fixed number of clusters. It doesn’t rely on labeled outcomes and doesn’t produce a label or a numeric prediction for each point. Instead, it discovers structure in the data by grouping together points that are close to each other in feature space, using cluster centroids to define those groups. That behavior aligns with clustering, not with predicting a class or a numerical target.

To contrast briefly: regression predicts a continuous value; classification assigns discrete labels based on learned boundaries from labeled data; binary and multiclass classifications are specific supervised forms of classification. Since K-Means learns without labels and aims to group similar items, clustering is the appropriate description. (Note: clustering can be used as a preprocessing step before classification in some workflows, but by itself it remains clustering.)