Description Usage Arguments Details Value Examples
Numerical optimization of an objective function G is carried out to find
appropriate signaldependent smoothing levels (λ's). This is easier
than visual inspection via the signaldependent tapering function in TaperingPlot
.
1 
Xmu 
Posterior mean of the input object as a vector. 
mm 
Number of rows of the original input object. 
nn 
Number of columns of the original input object. 
nGrid 
Size of grid where objective function is evaluated (nGridbynGrid).
This argument is ignorded if a sequence 
nLambda 
Number of lambdas to minimize over. Possible arguments: 2 (default) or 3. 
lambda 
λsequence which is used for optimization. If nothing is provided, 
sphere 

As signaldependent tapering functions are quiet irregular, it is hard to find appropriate smoothing values only by visual inspection of the tapering function plot. A more formal approach is the numerical optimization of an objective function.
Optimization can be carried out with 2 or 3 smoothing parameters. As the smoothing parameters 0 and ∞ are always added, this results in a mrbsizeR analysis with 4 or 5 smoothing parameters.
Sometimes, not all features of the input object can be extracted using the
smoothing levels proposed by MinLambda
. It might then be necessary to
include additional smoothing levels.
plot.minLambda
creates a plot of the objective function G
on a grid. The minimum is indicated with a white point. The minimum values of
the λ's can be extracted from the output of MinLambda
,
see examples.
A list with 3 objects:
G
Value of objective function G.
lambda
Evaluated smoothing parameters λ.
minind
Index of minimal λ's. lambda
[minind
]
gives the minimal values.
1 2 3 4 5 6 7 8 9 10 11  # Artificial sample data
set.seed(987)
sampleData < matrix(stats::rnorm(100), nrow = 10)
sampleData[4:6, 6:8] < sampleData[4:6, 6:8] + 5
# Minimization of two lambdas on a 20by20grid
minlamOut < MinLambda(Xmu = c(sampleData), mm = 10, nn = 10,
nGrid = 20, nLambda = 2)
# Minimal lambda values
minlamOut$lambda[minlamOut$minind]

Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.