Clustering and Dimensionality Reduction Quiz

What is the primary purpose of clustering in data analysis?

• To predict future outcomes based on existing data

• To find meaningful groupings in unlabelled data

• To create a labeled dataset for training

• To increase the speed of data processing

Which of the following methods is an example of a density-based clustering algorithm?

• DBSCAN

• K-means

• STING

• CURE

Why is hierarchical clustering considered a better choice compared to partitioning clustering for some datasets?

• It uses complex predictive models for clustering

• It mainly focuses on finding circular clusters

• It does not require pre-specifying the number of clusters

• It processes data much faster than other methods

Which of these is NOT an example of a partitioning method used in clustering?

• CLARANS

• Wave cluster

• K-means

• CURE

In which field is clustering NOT typically used based on the applications provided?

• Medical diagnosis

• Social network analysis

• Traffic analysis

• Forecasting weather conditions

Which characteristic of Hierarchical Divisive clustering makes it more complex compared to agglomerative clustering?

• It requires a flat clustering method as a subroutine to split each cluster.

• It merges clusters with random methods without prior distance measurement.

• It requires manual selection of the number of clusters at the start.

• It makes clustering decisions based solely on local patterns.

What is a key difference between K-Means and Hierarchical Divisive clustering regarding initial input requirements?

• Both methods require equal computation times by default.

• Hierarchical Divisive requires estimating the cluster size initially.

• K-Means requires the number of clusters to be predefined.

• K-Means groups data without user-defined centroids.

How does K-Means clustering determine cluster membership during iterations?

• By random assignment and no centroid recalibration each iteration.

• By assigning points to the closest centroid and recalculating centroids.

• By assigning points to the farthest centroid and fixing centroids initially.

• By merging clusters based solely on proximity to other clusters.

Which problem is DBSCAN particularly designed to overcome compared to traditional K-Means clustering?

• Employing supervised learning techniques for cluster validation.

• Handling clusters of arbitrary shapes without requiring a predefined number of clusters.

• Ensuring centroids have zero variance in distance to points.

• Avoiding all sensitivity to initial data ordering.

What limitation does K-Means clustering face when dealing with high-dimensional data?

• Difficulty in identifying meaningful clusters due to complex data geometry.

• Increased accuracy with use of non-Euclidean distances.

• Independence from initial centroid positions in high dimensions.

• Improved cluster tightness increases with dimensionality.