Introduction to Divisive Clustering Answers 1. What is a key characteristic of divisive clustering in contrast to agglomerative clustering? A. It does not utilize any distance-based criteria. B. It operates more efficiently on larger datasets. C. It begins with a single cluster and splits into smaller ones. D. It builds clusters by combining individual data points. Answer: It begins with a single cluster and splits into smaller ones. (C) Divisive clustering initiates the process with one encompassing cluster which is then divided. 2. Which criterion is used to assess the optimal division points in variance-based clustering? A. Maximized distance between points. B. Minimized variance among data points. C. Maximum density of data in clusters. D. Average distance from a centroid. Answer: Minimized variance among data points. (B) Variance-based criteria aim to minimize the spread of data points, hence minimizing variance. 3. What is one of the primary applications of divisive clustering in marketing? A. Segmenting customers based on purchasing behavior. B. Linking unrelated documents together. C. Grouping documents based on their length. D. Creating visual features in images. Answer: Segmenting customers based on purchasing behavior. (A) Divisive clustering can effectively group customers to tailor marketing efforts. 4. Which statement about sensitivity to initial conditions in clustering is true? A. It ensures consistent clustering results every time. B. It can lead to varied outcomes based on the optimization criterion. C. It is a non-factor in the clustering decision process. D. It has no impact on the clustering process. Answer: It can lead to varied outcomes based on the optimization criterion. (B) The selection of an optimization criterion significantly affects the clustering results. 5. In what scenario might divisive clustering be preferred over agglomerative clustering? A. When a distance-based criterion is essential. B. When exact cluster shapes are not crucial. C. When data points are uniformly distributed. D. When handling very large datasets. Answer: When a distance-based criterion is essential. (A) Divisive clustering is often better suited for identifying distinct clusters using distance-based criteria. 6. What is the primary starting point for divisive clustering? A. A single cluster containing all data points B. Individual data points as separate clusters C. Clusters determined by a predefined criterion D. Multiple clusters formed randomly Answer: A single cluster containing all data points (A) Divisive clustering begins with one cluster that encompasses all data points. 7. Which method relies on splitting clusters based on distances between their centroids? A. Agglomerative divisive clustering B. Hierarchical clustering C. Centroid-based divisive clustering D. Random clustering Answer: Centroid-based divisive clustering (C) Centroid-based divisive clustering focuses on the average distance between data samples and cluster centers. 8. What is one significant disadvantage of divisive clustering? A. It creates clusters incrementally B. High computational complexity, particularly with large datasets C. It provides visual representations of clusters D. It requires identifying potential division points Answer: High computational complexity, particularly with large datasets (B) The complexity of the algorithm can become a challenge, especially with large datasets. 9. Which of the following best describes the process of the divisive clustering algorithm? A. Randomly assigning data points to clusters B. Merging clusters based on proximity C. Iteratively dividing clusters until each data point forms its own cluster D. Establishing a single cluster without division Answer: Iteratively dividing clusters until each data point forms its own cluster (C) Divisive clustering iteratively divides clusters until every data point is its own cluster. 10. What type of hierarchical structure does divisive clustering produce? A. Clusters that are randomly generated B. A hierarchical representation of relationships between different clusters C. Flat clusters with no relationship D. Unrelated clusters based solely on data types Answer: A hierarchical representation of relationships between different clusters (B) Divisive clustering provides a hierarchical representation, allowing for easy examination of cluster relationships. 11. Which option describes the method's approach to identifying clusters during the algorithm? A. Removing outliers before clustering B. Assigning points to clusters based on color coding C. Evaluating the quality of each potential division based on data point distances D. Using random numbers to split Answer: Evaluating the quality of each potential division based on data point distances (C) The algorithm evaluates the quality of divisions by examining the distances among the data points. 12. Which type of clustering begins with each point in its own cluster before merging clusters? A. Agglomerative divisive clustering B. Random clustering C. Divisive hierarchical clustering D. Centroid-based divisive clustering Answer: Agglomerative divisive clustering (A) Agglomerative divisive clustering starts with each data point as its own cluster, merging them incrementally. 13. What benefit does divisive clustering provide in terms of data types? A. It automatically detects outliers B. It offers flexibility in selecting a criterion suitable for various data types C. It requires all data to be normalized D. It fixes the number of clusters in advance Answer: It offers flexibility in selecting a criterion suitable for various data types (B) Divisive clustering allows for flexibility in choosing the splitting criterion, accommodating different data types.