Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1146333 | Journal of Multivariate Analysis | 2012 | 11 Pages |
Abstract
We design a data-dependent metric in RdRd and use it to define the kk-nearest neighbors of a given point. Our metric is invariant under all affine transformations. We show that, with this metric, the standard kk-nearest neighbor regression estimate is asymptotically consistent under the usual conditions on kk, and minimal requirements on the input data.
Related Topics
Physical Sciences and Engineering
Mathematics
Numerical Analysis
Authors
Gérard Biau, Luc Devroye, Vida Dujmović, Adam Krzyżak,