Rearrange the following steps for K-means based segmentation: 1. Assign each observation to the group that has the closest centroid. 2. Repeat until either the centroid position or the observation assignments no longer move. 3. Guess K centroid positions, a centroid for each resulting cluster. 4. When all observations have been assigned, recalculate the positions of the K centroids.