Clustering and partitioning in dbt - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

What is the main goal of clustering in machine learning?

Clustering groups data points based on their similarities. What is the primary purpose of clustering?

A To reduce the number of features in the dataset

B To predict the output for new data points

C To group similar data points without using labeled data

D To split data into training and testing sets

💡 Hint

Think about whether clustering uses labels or not.

Clustering is an unsupervised learning method that groups data points based on similarity without using labels.

❓ Predict Output

intermediate

What is the output of this K-means clustering code snippet?

Given the following Python code using scikit-learn, what is the predicted cluster label for the point [1, 2]?

dbt

from sklearn.cluster import KMeans
import numpy as np

X = np.array([[1, 2], [1, 4], [1, 0], [10, 2], [10, 4], [10, 0]])
kmeans = KMeans(n_clusters=2, random_state=0).fit(X)
pred = kmeans.predict([[1, 2]])
print(pred[0])

A 1

B 0

C 2

D Error: n_clusters must be less than or equal to number of samples

💡 Hint

Look at how the data points are grouped and which cluster [1, 2] is closer to.

The points near [1, 2] form one cluster labeled 0, while points near [10, ...] form cluster 1.

❓ Model Choice

advanced

Which clustering algorithm is best for detecting clusters of varying shapes?

You have a dataset with clusters that are not spherical but have irregular shapes. Which algorithm is most suitable?

A DBSCAN clustering

B K-means clustering

C Hierarchical clustering with single linkage

D Linear regression

💡 Hint

Consider algorithms that do not assume cluster shape.

DBSCAN can find clusters of arbitrary shape by grouping dense regions, unlike K-means which assumes spherical clusters.

❓ Hyperparameter

advanced

What effect does increasing the number of clusters (k) have in K-means?

In K-means clustering, what happens if you increase the number of clusters k too much?

A Clusters become too general and lose detail

B The algorithm runs faster

C The model automatically finds the best k

D Clusters become smaller and may overfit the data

💡 Hint

Think about what happens when you split data into many small groups.

Increasing k creates many small clusters that may capture noise, leading to overfitting.

❓ Metrics

expert

Which metric evaluates clustering quality without true labels?

You want to measure how well your clustering algorithm performed but you do not have true labels. Which metric can you use?

A Silhouette score

B Accuracy

C Mean squared error

D Confusion matrix

💡 Hint

Look for a metric that works without knowing the correct groups.

Silhouette score measures how similar points are within clusters compared to other clusters without needing labels.