PeakPurity {ImpuR} | R Documentation |
Calculate a bilinear model and estimate noise from a spectime object and perform window evolving factor analysis.
PeakPurity(object, hetero = NULL, wefa.margin = 10, bilinear.model = "svd", err.win = 1:20, svd.nsvd = 4, peel.specfirst = FALSE, peel.direction = rep(c("left", "right"), 2), favor.positive = FALSE, verbose = FALSE)
object |
a spectime object |
hetero |
whether to do heteroscedasticity correction in wefa |
wefa.margin |
window width for wefa |
bilinear.model |
which bilinear model to use (svd or peel) |
err.win |
times (rows) on which to estimage noise |
svd.nsvd |
number of singular components in svd |
peel.specfirst |
peeling: whether to peel spectrum first |
peel.direction |
peeling: come from "left" (top) or "right" (bottom) |
favor.positive |
whether to favor positive components |
verbose |
echo stages |
Use ErrorComponents to calculate noise. Then use Globalsvd or Peel to extract bilinear model. The carry out window evolving factor analysis.
a spectime
object for which the slots @std.err, @hetero, @ct,
@spec, @WEAtime, @WEAwavelength are filled in.
Christian Ritter
Ritter C. and Gilliard J. (2007), JSS
data(al1) plot(PeakPurity(al1,bilinear.model="svd")) plot(PeakPurity(al1,bilinear.model="peel"))