checkfit: Check goodness-of-fit of GaGa and MiGaGa models
In gaga: GaGa hierarchical model for high-throughput data analysis

Description Usage Arguments Details Value Note Author(s) References See Also

Produces plots to check fit of GaGa and MiGaGa model. Compares observed data with posterior predictive distribution of the model. Can also compare posterior distribution of parameters with method of moments estimates.

1	checkfit(gg.fit, x, groups, type='data', logexpr=FALSE, xlab, ylab, main, lty, lwd, ...)

`gg.fit`	GaGa or MiGaGa fit (object of type `gagafit`, as returned by `fitGG`).
`x`	`ExpressionSet`, `exprSet`, data frame or matrix containing the gene expression measurements used to fit the model.
`groups`	If `x` is of type `ExpressionSet` or `exprSet`, `groups` should be the name of the column in `pData(x)` with the groups that one wishes to compare. If `x` is a matrix or a data frame, `groups` should be a vector indicating to which group each column in x corresponds to.
`type`	`data` checks marginal density of the data; `shape` checks shape parameter; `mean` checks mean parameter; `shapemean` checks the joint of shape and mean parameters
`logexpr`	If set to `TRUE`, the expression values are in log2 scale.
`xlab`	Passed on to `plot`
`ylab`	Passed on to `plot`
`main`	Passed on to `plot`
`lty`	Ignored.
`lwd`	Ignored.
`...`	Other arguments to be passed to `plot`

The routine generates random draws from the posterior and posterior predictive distributions, fixing the hyper-parameters at their estimated value (posterior mean if model was fit with method=='Bayes' or maximum likelihood estimate is model was fit with method=='EBayes').

Produces a plot.

Posterior and posterior predictive checks can lack sensitivity to detect model misfit, since they are susceptible to over-fitting. An alternative is to perform prior predictive checks by generating parameters and data with simGG.

David Rossell

Rossell D. GaGa: a simple and flexible hierarchical model for microarray data analysis. http://rosselldavid.googlepages.com.

simGG to simulate samples from the prior-predictive distribution, simnewsamples to generate parameters and observations from the posterior predictive, which is useful to check goodness-of-fit individually a desired gene.

gaga documentation built on Nov. 8, 2020, 5:49 p.m.