<p>either "gaussian" or "epanechnikov"</p>

type

internal

Estimates conditional distributions and conditional quantiles.  The versions of the methods in this package are primarily for use in multiple step procedures where the first step is to estimate a conditional distribution.  In particular, there are functions for implementing distribution regression, quantile regression, and versions of local linear distribution regression; all in a unified framework.  Distribution regression provides a way to flexibly model the distribution of some outcome Y conditional on covariates X without imposing parametric assumptions on the conditional distribution but providing more structure than fully nonparametric estimation (See Foresi and Peracchi (1995) <doi:10.2307/2291056> and Chernozhukov, Fernandez-Val, and Melly (2013) <doi:10.3982/ECTA10582>).

k: k

Description

Usage

Arguments

Value