What differentiates agglomerative clustering from divisive clustering?

Agglomerative clustering and divisive clustering are both hierarchical clustering methods but approach the clustering process from different directions.

Agglomerative clustering is characterized by its bottom-up approach. It starts with each individual data point as its own separate cluster. Then, it iteratively merges the closest clusters together until all points are combined into a single cluster or until a specified number of clusters is reached. This method effectively builds larger clusters from smaller ones, making it intuitive for understanding how clusters form from individual observations.

In contrast, divisive clustering takes a top-down approach. It begins with all data points grouped into one single cluster and then recursively splits that cluster into smaller sub-clusters.

Understanding this fundamental difference highlights why the correct choice indicates that agglomerative clustering starts with each example as its own cluster, as it encapsulates the essence of the method itself.