Skip to yearly menu bar Skip to main content


Finite-Sample Analysis of Fixed-k Nearest Neighbor Density Functional Estimators

Shashank Singh · Barnabas Poczos

Area 5+6+7+8 #48

Keywords: [ Learning Theory ] [ (Other) Statistics ] [ Information Theory ]


We provide finite-sample analysis of a general framework for using k-nearest neighbor statistics to estimate functionals of a nonparametric continuous probability density, including entropies and divergences. Rather than plugging a consistent density estimate (which requires k → ∞ as the sample size n → ∞) into the functional of interest, the estimators we consider fix k and perform a bias correction. This can be more efficient computationally, and, as we show, statistically, leading to faster convergence rates. Our framework unifies several previous estimators, for most of which ours are the first finite sample guarantees.

Live content is unavailable. Log in and register to view live content