loessPred: Fit a loess model to the p-value data for later use in...
In dalejbarr/simgen: Monte Carlo Simulations of Simultaneous Generalization

Description Usage Arguments Value See Also Examples

Given a vector of p-values and two population parameters, this will fit a loess model to the data, with the model of the form z ~ x * y, where z is the predicted p-value. This can later be used to make a heatmap, and is separated from the heatmap2 function because of the long amount of time it can take for this function to run.

1
2
3

loessPred(pval, paramX, paramY, minx = 0.01, maxx = 2.99,
  miny = 0.01, maxy = 2.99, svec = rep(TRUE, length(paramX)),
  span = 0.9)

`pval`	vector of p-values from simulated runs
`paramX`	population parameter to condition `z` on (e.g., a column from a matrix created using `createParamMx`). This will appear on the x-axis of the heatmap.
`paramY`	same as `paramX`, but for the Y-axis
`minx`	minimum x value
`maxx`	maximum x value
`miny`	minimum y value
`maxy`	maximum y value
`svec`	a logical vector to select out only certain rows from `pval`, `paramX` and `paramY`
`span`	a value to be passed along to the `loess` function that determines the smoothing level

A fitted loess model.

loess, heatmap2

pmx <- createParamMx(100)

result <- apply(pmx, 1, function(x) {fitanova(mkDf(24,24,mcr.params=x,FALSE),FALSE)})
lpred <- loessPred(pval=result["pmax",],
                paramX=pmx[,"t11"], paramY=pmx[,"w11"])

heatmap2(lpred, "t11", "w11", "F1+F2")