Challenge - 5 Problems

🎖️

Dimensionality Reduction Master

Get all challenges correct to earn this badge!

Test your skills under time pressure!

🧠 Conceptual

intermediate

2:00remaining

Purpose of Dimensionality Reduction Visualization

Why do we use dimensionality reduction techniques like t-SNE or PCA for visualization in machine learning?

ATo reduce the number of features so the model trains faster without losing any information

BTo increase the number of features for better model accuracy

CTo transform high-dimensional data into 2D or 3D so humans can visually understand patterns

DTo remove noisy data points from the dataset automatically

Attempts:

2 left

❓ Predict Output

intermediate

2:00remaining

Output of PCA Dimensionality Reduction Code

What is the shape of the transformed data after applying PCA with 2 components on a dataset with shape (100, 5)?

ML Python

from sklearn.decomposition import PCA
import numpy as np
X = np.random.rand(100, 5)
pca = PCA(n_components=2)
X_pca = pca.fit_transform(X)
print(X_pca.shape)

A(100, 2)

B(2, 100)

C(5, 2)

D(100, 5)

Attempts:

2 left

❓ Model Choice

advanced

2:00remaining

Choosing a Dimensionality Reduction Method for Visualization

You have a dataset with 1000 features and want to visualize clusters in 2D. Which method is best for preserving local structure and showing clusters clearly?

APCA because it preserves global variance and is fast

BLinear Regression because it reduces features by fitting a line

CK-Means clustering because it reduces dimensions by grouping data

Dt-SNE because it preserves local neighbor distances and reveals clusters well

Attempts:

2 left

❓ Metrics

advanced

2:00remaining

Interpreting Explained Variance Ratio in PCA

After applying PCA with 3 components, the explained variance ratios are [0.5, 0.3, 0.1]. What does this tell you about the data?

AThe components explain 50%, 30%, and 10% of the samples respectively

BThe first component explains 50% of the variance, the second 30%, and the third 10%, totaling 90% variance explained

CThe PCA model is overfitting because variance ratios should sum to 1

DThe data has 3 features only because PCA returned 3 components

Attempts:

2 left

🔧 Debug

expert

3:00remaining

Debugging t-SNE Visualization Code

You run this code to visualize data with t-SNE but get a ValueError: 'perplexity must be less than n_samples'. What is the cause?

ML Python

from sklearn.manifold import TSNE
import numpy as np
X = np.random.rand(10, 50)
tsne = TSNE(n_components=2, perplexity=30)
X_embedded = tsne.fit_transform(X)
print(X_embedded.shape)

AThe perplexity value 30 is too high for only 10 samples; it must be less than the number of samples

BThe number of components must be equal to the number of features

CThe input data X must be 2D but here it is 3D

DTSNE requires the input data to be normalized between 0 and 1

Attempts:

2 left