gauss_loo {mig}R Documentation

Leave-one-out cross-validation for Gaussian kernel density estimation

Description

Given a data matrix, compute the log density using leave-one-out cross validation, taking in turn an observation as location vector and computing the density of the resulting mixture.

Usage

gauss_loo(x, Sigma, logweights)

Arguments

x

n by d matrix of observations

Sigma

smoothing positive-definite matrix

logweights

vector of log weights

Value

a vector of values for the weighted leave-one-out likelihood


[Package mig version 2.0 Index]