To prevent a k-Means model for a large data set from taking a long time to run, you can adjust the _________ parameter in RapidMiner.

• A. Max iterations
• B. Max runs
• C. Learning rate
• D. Number of clusters

Answer: (B)

When clustering with k-Means, you typically run the algorithm multiple times with different starting centroids to avoid poor local minima. Each run goes through the data, updating centroids until convergence, so the total time grows with the number of restarts. By lowering the maximum number of runs in RapidMiner, you reduce the total amount of work the algorithm does, speeding up the process on large data sets. This parameter directly controls how many full explorations of centroid positions are performed, which is why it’s the best lever for speeding up training. Other options affect time in different ways: max iterations changes work within a single run, the learning rate isn’t used in basic k-Means, and the number of clusters changes the cost per iteration but isn’t a direct global speed control.