On the distance between mean and geometric median in high dimensions
By: Richard Schwank, Mathias Drton
Potential Business Impact:
Makes computer guesses more accurate with more data.
The geometric median, a notion of center for multivariate distributions, has gained recent attention in robust statistics and machine learning. Although conceptually distinct from the mean (i.e., expectation), we demonstrate that both are very close in high dimensions when the dependence between the distribution components is suitably controlled. Concretely, we find an upper bound on the distance that vanishes with the dimension asymptotically, and derive a rate-matching first order expansion of the geometric median components. Simulations illustrate and confirm our results.
Similar Papers
Geometric medians on product manifolds
Methodology
Finds the best average shape from many different kinds.
Performance of the empirical median for location estimation in heteroscedastic settings
Statistics Theory
Finds the true middle point in messy data.
The empirical median for estimating the common mean of heteroscedastic random variables
Statistics Theory
Finds the average of numbers that are spread out.