Boost K-Means

Due to its simplicity and versatility, k-means remains popular since it was proposed three decades ago. Since then, continuous efforts have been taken to enhance its performance. Unfortunately, a good trade-off between quality and efficiency is hardly reached. In this paper, a novel k-means variant is presented. Different from most of k-means variants, the clustering procedure is explicitly driven by an objective function, which is feasible for the whole l2-space. The classic egg-chicken loop in k-means has been simplified to a pure stochastic optimization procedure. K-means therefore becomes simpler, faster and better. The effectiveness of this new variant has been studied extensively in different contexts, such as document clustering, nearest neighbor search and image clustering. Superior performance is observed across different scenarios.