KMeans

KMeans is an efficient, unsupervised clustering algorithm that partitions $n$ observations into $k$ distinct clusters, minimizing the within-cluster sum-of-squares (inertia).

This algorithm operates iteratively: first, it selects $k$ initial centroids (cluster centers), often using the `k-means++` method for better initialization. Next, it assigns every data point to the nearest centroid, typically using Euclidean distance. The third step recalculates the new centroid as the mean of all points assigned to that cluster. The process repeats—assigning points, then recalculating centroids—until the centroids stabilize or a maximum iteration count (e.g., 300) is reached. The core objective is to achieve a locally optimal solution that minimizes the distance between data points and their assigned cluster center, making it highly effective for tasks like customer segmentation and vector quantization.

https://scikit-learn.org/stable/modules/clustering.html#k-means