global-optimization

Proposes Big-means++, a simple algorithm that achieves global optimization quality for big data K-means clustering by systematically curating inputs and using sample-induced surrogate landscapes.