_find_closest_centroid (
        vector,
        centroids,
        distance_fn,
        )

_find_closest_centroid(vector, centroids, distance_fn) -> index of closest centroid

cluster (
        data,
        k,
        distance_fn=distance.euclidean,
        init_centroids_fn=random_centroids,
        max_iterations=1000,
        update_fn=None,
        )

cluster(data, k[, distance_fn][, max_iterations][, update_fn]) -> (centroids, clusters) or None

Organize data into k clusters. Return a list of cluster assignments between 0-(k-1), where the items in the list corresponds to the list of data points. If the algorithm does not converge by max_iterations (default is 1000), returns None. data is a list of data points, which are vectors of numbers. distance_fn is a callback function that calculates the distance between two vectors. By default, the Euclidean distance wwill be used. If update_fn is specified, it is called at the beginning of every iteration and passed the iteration number, cluster centroids, and current cluster assignments.

first_k_points_as_centroids ( data,  k )

first_k_points_as_centroids(data, k) -> list of centroids

Picks the first K points as the initial centroids. This isn't a good method (unless the data is randomized), but does provide determinism that's useful for debugging.

random_centroids ( data,  k )

random_centroids(data, k) -> list of centroids

Return a list of data points to serve as the initial centroids. This is k randomly chosen data points. Tries to avoid having repeated centroies, if possible.