numEff {bayesm}R Documentation

Compute Numerical Standard Error and Relative Numerical Efficiency

Description

numEff computes the numerical standard error for the mean of a vector of draws as well as the relative numerical efficiency (ratio of variance of mean of this time series process relative to iid sequence).

Usage

numEff(x, m = as.integer(min(length(x),(100/sqrt(5000))*sqrt(length(x)))))

Arguments

x

R x 1 vector of draws

m

number of lags for autocorrelations

Details

default for number of lags is chosen so that if R=5000, m=100 and increases as the sqrt(R).

Value

A list containing:

stderr

standard error of the mean of x

f

variance ratio (relative numerical efficiency)

Warning

This routine is a utility routine that does not check the input arguments for proper dimensions and type.

Author(s)

Peter Rossi, Anderson School, UCLA, perossichi@gmail.com.

References

For further discussion, see Chapter 3, Bayesian Statistics and Marketing by Rossi, Allenby, and McCulloch.
http://www.perossi.org/home/bsm-1

Examples

numEff(rnorm(1000), m=20)
numEff(rnorm(1000))

[Package bayesm version 3.1-4 Index]