Smooth projected density estimation

TL;DR

The paper introduces a new nonparametric density estimator called the smooth projection estimator (SPE), which uses a least squares projection onto a mixture class with an undersmoothed pilot estimate to achieve optimal convergence rates.

Contribution

It presents the SPE, a novel density estimation method that combines convex analysis and mixture densities for improved efficiency and convergence in multi-dimensional settings.

Findings

SPE achieves optimal convergence rates in $ ext{L}_2$ norm.

The solution dimension is at most $n+1$, enhancing computational efficiency.

SPE is easy to compute, store, and evaluate at new data points.

Abstract

We introduce and analyse a new nonparametric estimator of a multi-dimensional density. Our smooth projection estimator (SPE) is defined by a least squares projection of the sample onto an infinite dimensional mixture class via an undersmoothed nonparametric pilot estimate, which acts as a structural filter to regularise the solution. The undersmoothing is required to optimise the convergence rate of the SPE, which is jointly determined by that of the pilot estimator to the true density in squared $\mathbb{L}_{2}$ norm, and by that of the pilot distribution function to the empirical distribution function in uniform norm. Our procedure was conceived with a view to exploiting well known results in convex analysis and their connection to mixture densities. In the context of our work, this translates to the observation that the infinite dimensional minimisation problem, implicit in the construction of the SPE, possesses a solution of dimension at most $n+1$, where $n$ is the sample size. The SPE thus enjoys practical advantages such as computational efficiency, ease of storage and rapid evaluation at a new data point.