Novel adaptive family of partitioning algorithms

In this talk, I will present a unified family of data-adaptive partitioning algorithms that extends several well-known methods (e.g., k-means and k-subspaces). Indexed by a single parameter and employing a common centroidal Voronoi tessellation-based minimization strategy, the algorithms are easy to use and interpret, and scale well to large, high-dimensional problems. The data-adaptive framework developed in this work: (a) exhibits skill at automatically uncovering data structures and problem parameters without any expert knowledge and, (b) can be used to augment other existing methods. Analytical results and numerical experiments presented in this talk will highlight advantages of the new methodology in the context of several disparate application areas including subspace clustering, model order reduction, and matrix approximation. 

Bio
https://math.gmu.edu/~memelian/