Tuning parameters for the algorithms used by function ojaMedian.
ojaMedianControl(sigmaInit = 0, sigmaAda = 20, adaFactor = 0.5,
iter = 1e+06, useAllSubsets = FALSE,
nSubsetsUsed = 1000, sigmaLog10Dec = 10,
storeSubDet = TRUE, eps = 0.1, chi2 = 0.95,
samples = 20, maxlines = 1000, S1 = cov,
S2 = cov4, S1args = list(), S2args = list(),
volume = 1e-6, boundedExact = T, ...)
(for the evo
algorithm): Set the initial variance of the mutation vector in the first run.
(for the evo
algorithm): Defines after how many mutations the variance of the mutation vector is adjusted.
(for the evo
algorithm): Defines the level of adjustment of the mutation vector.
(for the evo
algorithm): The maximum number of iterations. If the algorithm does not converge, it stops after iter
- iterations.
(for the evo
algorithm): A logical flag. If it is set all datapoints and resulting simplices are taken into account for the calculation.
(for the evo
algorithm): If useAllSubsets is not set, this determines how many, randomly selected, datapoints are taken into account.
(for the evo
algorithm): This is an abort criterion. If the logarithmised initial variance differs more than sigmaLog10Dec from the actual, logarithmised variance, the algorithm stops.
(for the evo
algorithm): A boolean flag. If it is set subdeterminants are stored. This should always been set to TRUE
if \(6*(dim -1)*nSubsetsUsed<CPU-Cache\).
(for the grid
algorithm): This is the abort criterion. If the grid becomes denser than this threshold the algorithm stops.
(for the grid
algorithm): This is the test niveau of the test, if a grid point could be used as a Oja-Median or not.
(for the grid
algorithm): This determines how many additional hyperplanes are taken after every run.
(for the exact
algorithm): This determines how many intersection lines are investigated in addtion to the one with the steepest gradient.
(for the evo
and grid
algorithms): Optional arguments for S1
passed on to ics
to compute the invariant coordinate system.
(for the evo
and grid
algorithms): Optional arguments for S2
passed on to ics
to compute the invariant coordinate system.
(for the bounded_exact
algorithm): is the desired size of the bounded region, which is selected as a part of the original volume. Here the volume is calculated as the volume of a minimal multivariate circumscribed rectangle with edges parallel to the coordinate axes.
(for the bounded_exact
algorithm): setting this parameter to FALSE
stops the algorithm after the bounded region is found, and its center is reported as an approximation of the median.
Additional hidden option
ojaMedian
, also for references and examples.
# NOT RUN {
## Show the default settings:
str(ojaMedianControl())
# }
Run the code above in your browser using DataLab