Unsupervised Learning DY0-001 Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A marketing analyst wants to segment customers into groups with similar purchasing behavior. They have data on annual spending and frequency of purchases. They are unsure how many segments exist. Which approach should they use to determine the optimal number of clusters?

   • AUse SVD to find the number of singular values
   • BRun K-Means with k=2 and k=3 and pick the one with lower inertia
   • CRun K-Means for a range of k values and use the elbow method
   • DUse PCA to determine the number of clusters
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A streaming service wants to build a recommendation system that predicts user ratings for movies. They have a large user-item rating matrix with many missing values. Which technique is most appropriate for this task?

   • AHierarchical clustering on movies
   • BK-Means clustering on users
   • CPCA to reduce the item features
   • DSingular Value Decomposition (SVD) for matrix factorization
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A data engineer is preparing a dataset with 1,000 features for a supervised learning model. The model training is slow and suffers from the curse of dimensionality. They want to reduce the number of features while preserving as much information as possible. Which technique should they apply?

   • ASVD for matrix factorization
   • BPrincipal Component Analysis (PCA)
   • CHierarchical clustering
   • DK-Means clustering
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A data scientist has a dataset with 10,000 features and wants to reduce dimensionality for a clustering task. They are considering PCA and SVD. They need to choose the method that is most computationally efficient for this large feature space. Which method should they choose?

   • AHierarchical clustering
   • BSVD on the data matrix
   • CPCA via eigendecomposition of the covariance matrix
   • DK-Means clustering
   Reveal Solution Discussion Select an answer first
6. 20expert · hard

   A data scientist is using K-Means to segment customers. They notice that the clusters are not well-separated and the results vary significantly between runs. They suspect the issue is due to the initial placement of centroids. Which approach should they take to improve the stability and quality of the clustering?

   • AIncrease the number of clusters
   • BUse K-Means++ for initialization
   • CUse hierarchical clustering instead
   • DUse PCA to reduce to 2 dimensions
   Reveal Solution Discussion Select an answer first