bicTuning: Tuning Parameter Selection via the Bayesian Information...

In tmanole/GroupSortFuse: Estimating the Number of Components in Finite Mixture Models

Description

Bayesian Information Criterion (BIC) for tuning parameter selection.

Usage

bicTuning(y, result)

Arguments

y	n by D matrix consisting of the data.
result	A gsf object.

Details

The BIC selects the best tuning parameter out of the ones used in a gsf object by minimizing the following function

\textrm{BIC}(λ) = -2 l_n(\hat{\mathbf{Ψ}}_{λ}) + \textrm{df}(\hat{\mathbf{Ψ}}_{λ}) \log n

where l_n is the log-likelihood function, and \hat{\mathbf{Ψ}}_{λ} is the set of estimated parameters theta and pii corresponding to the tuning parameter λ. \textrm{df}(\hat{\mathbf{Ψ}}_{λ}) denotes the degrees of freedom of the estimates.

Value

A list containing the selected tuning parameter and corresponding estimates, as well as a summary of the computed RBIC values, log-likelihood values, and corresponding orders of the estimates.

Examples

 require(MM)
 data(pollen)
 set.seed(1)
 out <- multinomialOrder(pollen, K=12, lambdas=seq(0.1, 1.6, 0.2))
 tuning <- bicTuning(pollen, out)
 plot(out, eta=TRUE, gg=FALSE, opt=tuning$result$lambda)

tmanole/GroupSortFuse documentation built on Jan. 12, 2022, 10:37 p.m.