42
13
v1v2 (latest)

Achieving the time of 11-NN, but the accuracy of kk-NN

Abstract

We propose a simple approach which, given distributed computing resources, can nearly achieve the accuracy of kk-NN prediction, while matching (or improving) the faster prediction time of 11-NN. The approach consists of aggregating denoised 11-NN predictors over a small number of distributed subsamples. We show, both theoretically and experimentally, that small subsample sizes suffice to attain similar performance as kk-NN, without sacrificing the computational efficiency of 11-NN.

View on arXiv
Comments on this paper