An adaptive multiclass nearest neighbor classifier

We consider a problem of multiclass classification, where the training sample is generated from the model , , and are unknown -Holder continuous functions.Given a test point , our goal is to predict its label. A widely used -nearest-neighbors classifier constructs estimates of and uses a plug-in rule for the prediction. However, it requires a proper choice of the smoothing parameter , which may become tricky in some situations. In our solution, we fix several integers , compute corresponding -nearest-neighbor estimates for each and each and apply an aggregation procedure. We study an algorithm, which constructs a convex combination of these estimates such that the aggregated estimate behaves approximately as well as an oracle choice. We also provide a non-asymptotic analysis of the procedure, prove its adaptation to the unknown smoothness parameter and to the margin and establish rates of convergence under mild assumptions.
View on arXiv