Virtual Labs

Clustering: K-Means Algorithm

Which statement best explains why K-Means uses squared Euclidean distance instead of simple Euclidean distance?

a: To reduce computational complexity Explanation

Explanation

b: To penalize larger deviations more strongly Explanation

Explanation

c: To normalize data automatically Explanation

Explanation

d: To make clusters non-spherical Explanation

Explanation

What is the main reason K-Means may produce different results on different runs?

a: Random initialization of centroids Explanation

Explanation

b: Different hardware performance Explanation

Explanation

c: Change in dataset size Explanation

Explanation

d: Deterministic nature of algorithm Explanation

Explanation

Which condition ensures that K-Means has converged?

a: Centroids stop moving Explanation

Explanation

b: Maximum iterations reached Explanation

Explanation

c: Distance becomes zero Explanation

Explanation

d: Clusters become equal in size Explanation

Explanation

The distance of a point in a cluster to another point in the same cluster is generally

a: Greater than its own cluster points Explanation

Explanation

b: Lesser than its own cluster points Explanation

Explanation

c: Equal Explanation

Explanation

d: None of above Explanation

Explanation

Why is feature scaling important in K-Means?

a: To reduce number of clusters Explanation

Explanation

b: To equalize influence of all features in distance calculation Explanation

Explanation

c: To increase accuracy automatically Explanation

Explanation

d: To remove outliers Explanation

Explanation

Which of the following best describes the role of centroids?

a: They act as boundary lines Explanation

Explanation

b: They represent cluster centers Explanation

Explanation

c: They store all data points Explanation

Explanation

d: They remove noise Explanation

Explanation

If clusters are highly overlapped, this indicates

a: Poor clustering Explanation

Explanation

b: Good clustering Explanation

Explanation

c: Average clustering Explanation

Explanation

d: None of above Explanation

Explanation

Why does K-Means prefer compact clusters?

a: Because it minimizes variance within clusters Explanation

Explanation

b: Because it maximizes entropy Explanation

Explanation

c: Because it uses probability models Explanation

Explanation

d: Because it maximizes cluster size Explanation

Explanation

Which property of K-Means makes it unsuitable for categorical data?

a: It uses probability Explanation

Explanation

b: It requires mean computation Explanation

Explanation

c: It depends on labeled data Explanation

Explanation

d: It uses density Explanation

Explanation

Why is K-Means not suitable for clusters with complex shapes

a: Because it assumes clusters are spherical around centroids Explanation

Explanation

b: Because it cannot compute distances Explanation

Explanation

c: Because it requires labeled data Explanation

Explanation

d: Because it cannot handle numerical data Explanation

Explanation