Local-Likelihood Transformation Kernel Density Estimation for Positive Random Variables

Abstract

The kernel estimator is known not to be adequate for estimating the density of a positive random variable X. The main reason is the well-known boundary bias problems that it suffers from, but also its poor behaviour in the long right tail that such a density typically exhibits. A natural approach to this problem is to first estimate the density of the logarithm of X, and obtaining an estimate of the density of X using standard results on functions of random variables ('back-transformation'). Although intuitive, the basic application of this idea yields very poor results, as was documented earlier in the literature. In this paper, the main reason for this underachievement is identified, and an easy fix is suggested. It is demonstrated that combining the transformation with local likelihood density estimation methods produces very good estimators of R + -supported densities, not only close to the boundary, but also in the right tail. The asymptotic properties of the proposed 'local likelihood transformation kernel density estimators' are derived for a generic transformation, not only for the logarithm, which allows one to consider other transformations as well. One of them, called the 'probex' transformation, is given more focus. Finally, the excellent behaviour of those estimators in practice is evidenced through a comprehensive simulation study and the analysis of several real data sets. A nice consequence of articulating the method around local-likelihood estimation is that the resulting density estimates are typically smooth and visually pleasant, without oversmoothing important features of the underlying density.

Assessing the Global and Local Uncertainty of Scientific Evidence in the Presence of Model Misspecification

“…We use the R package kde1d (version 1.0.2, Nagler and Vatter, 2019), which uses univariate local polynomial(log-quadratic) kernel density estimators. Our validation tests support the literature (Geenens and Wang, 2018) on the strength of this method. We find that confidence bounds are located more accurately with kde1d quantiles than with raw bootstrap quantiles, BCa quantiles, or with calibrated (double bootstrap) quantiles (see Efron and Tibshirani, 1993 for description of these methods) and that estimated distributions are more accurate (in integrated squared error) than standard kernel density estimation.…”

Early human impacts and ecosystem reorganization in southern-central Africa

“…For the present analysis, each age was replotted 10,000 times along a normal distribution using the rnorm command in R based on the laboratory generated mean and 1- SE. The KDE was created in the "kde1d" package in R (69). Bandwidth was set to default, with data-derived parameters developed by Sheather and Jones (70).…”

Location-Based Social Network’s Data Analysis and Spatio-Temporal Modeling for the Mega City of Shanghai, China

“…For example, a check-in with the latitude and longitude of a university belongs to the 'Educational' category, whereas a check-in location in different restaurants is in the 'Food' category. For the spatial analysis, we used ArcMap, the Kernel Density Estimation (KDE) technique [18], and geospatial data (Shape Files) from OpenStreetMap [19].…”