GVmix: NPMLE for Gaussian Variance Heterogeneity

Description

A Kiefer-Wolfowitz MLE for Gaussian models with independent variances. This can be viewed as a general form for χ^2 mixtures, see Gammamix for a more general form for Gamma mixtures.

Usage

1
GVmix(x, m, v = 300, weights = NULL, ...)

Arguments

x

vector of observed variances

m

vector of sample sizes corresponding to x

v

A vector of bin boundaries, if scalar then v equally spaced bins are constructed

weights

replicate weights for x obervations, should sum to 1

...

optional parameters passed to KWDual to control optimization

Value

An object of class density with components:

x

midpoints of the bin boundaries

y

estimated function values of the mixing density

g

function values of the mixture density at the observed x's.

logLik

the value of the log likelihood at the solution

dy

Bayes rule estimates of

status

the Mosek convergence status.

Author(s)

R. Koenker

References

Koenker, R and I. Mizera, (2013) “Convex Optimization, Shape Constraints, Compound Decisions, and Empirical Bayes Rules,” JASA, 109, 674–685.

Gu J. and R. Koenker (2014) Unobserved heterogeneity in income dynamics: an empirical Bayes perspective, JBES, forthcoming.

See Also

Gammamix for a general implementation for Gamma mixtures


Questions? Problems? Suggestions? or email at ian@mutexlabs.com.

All documentation is copyright its authors; we didn't write any of that.