What term describes the process of repeating allocation and recalculation until the centroids stop changing?

What term describes the process of repeating allocation and recalculation until the centroids stop changing?

• A. Convergence
• B. Stability
• C. Termination
• D. Plateau

Answer: (A)

Convergence describes the state reached in an iterative clustering routine when repeated allocation and recomputation of centroids lead to no further movement. In k-means, you assign points to the nearest centroid, recompute the centroids, and repeat. When the centroids stop changing between iterations (or only change by an amount below a small threshold), you’ve arrived at a fixed point—the algorithm has converged. This term cleanly captures the idea of the process settling down into a stable configuration. Stability isn’t the formal term for this stopping behavior, and termination describes the act of stopping rather than the idea of reaching a stable, unchanged state. Plateau is a casual way to describe a lack of progress, but convergence is the standard, precise term used in this context.

Convergence describes the state reached in an iterative clustering routine when repeated allocation and recomputation of centroids lead to no further movement. In k-means, you assign points to the nearest centroid, recompute the centroids, and repeat. When the centroids stop changing between iterations (or only change by an amount below a small threshold), you’ve arrived at a fixed point—the algorithm has converged. This term cleanly captures the idea of the process settling down into a stable configuration.

Stability isn’t the formal term for this stopping behavior, and termination describes the act of stopping rather than the idea of reaching a stable, unchanged state. Plateau is a casual way to describe a lack of progress, but convergence is the standard, precise term used in this context.