The popular k-means algorithm is used to discover clusters in vector data automatically. We present three accelerated algorithms that compute exactly the same clusters much faster than the standard method. First, we redesign ...