Description Usage Arguments Value Author(s) See Also Examples
View source: R/registerIterated.R
Uses the fda::register.fd
function to register a sample of curves, after smoothing them, iteratively. Checks for outliers before estimating the weighted average. The weights are estimated based on distance from a L1-median robustX::L1median
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 | registerIterated(
dataToRegister,
Lambdas_ConstrainedWarping = c(0.001, 0.0001, 0.00005),
LambdaDefault = 0.00001,
abscissaFrom,
abscissaTo,
abscissaIncrement,
basisOrder = 5,
basisBreakFreq = 3,
Lambdas_Roughness = exp(-5:5),
outlierTrimPct = 0.2,
N_Bootstrap_for_Outlier = 500,
RE_REGISTER = FALSE,
Convergence_Threshold = 1e-05,
SimMeanDiff_Threshold = 0.0001,
MinSimilarityThreshold = 0.15,
MAX_ITERATION = 6
)
|
dataToRegister |
A matrix of curves to register. Each column is a new curve. |
Lambdas_ConstrainedWarping |
The roughness penalty for estimating the warping function. Smaller values will allow more undulations in the warping function. |
LambdaDefault |
The default roughness penalty, set to 0.00001. The λ is set to this value if the function iterates longer than the list of λ provided in |
abscissaFrom |
Mininum value of abscissa |
abscissaTo |
Maximum value of abscissa |
abscissaIncrement |
Increment of abscissa |
basisOrder |
Order of B-spline basis functions |
basisBreakFreq |
Frequency of knots of basis functions |
Lambdas_Roughness |
Roughness penalty for smoothing using B-splines |
outlierTrimPct |
Percentage to trim when detecting outliers |
N_Bootstrap_for_Outlier |
Number of bootstrap samples to draw, to estimate outlier curves |
RE_REGISTER |
Re-register the curves from previous iterations, or register the original noisy curves, at every iteration, to the updated consensus/template |
Convergence_Threshold |
The value for argument |
SimMeanDiff_Threshold |
Defaults to 0.001, a criterion to stop the iterations |
MinSimilarityThreshold |
Minimum similarity between registered curve and template to estimate the final consensus |
MAX_ITERATION |
Maximum number of iterations |
dataToRegister |
Original dataset to register |
Regfd_Final |
Final registered data object. Output of |
registeredCurves |
A matrix of registered curves extracted from the registered object |
registeredCurves.D1 |
A matrix of first derivatives of registered curves extracted from the registered object |
FinalConsensus |
Final consensus of registered curves |
curvesForConsensus |
Names of curves used to estimate the consensus. This excludes the detected outliers |
registeredCurvesAll |
A matrix of all registered curves, including outliers |
registeredCurvesAll.D1 |
A matrix of first derivatives of all registered curves, including outliers |
Sim_toTemplate |
A matrix of similarities of curves to template |
Subhrangshu Nandi; snandi@wisc.edu or nands31@gmail.com
register.fd
, registerSingleIter
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 | data( growth, package = 'fda' )
Mat1 <- growth[['hgtm']]
Arguments <- growth[['age']]
#Mat1Registered <- registerIterated(
# dataToRegister = Mat1,
# Lambdas_ConstrainedWarping = c(0.001, 0.0001, 0.00005),
# LambdaDefault = 0.00001,
# abscissaFrom = 1,
# abscissaTo = 16,
# abscissaIncrement = 0.5,
# basisOrder = 5,
# basisBreakFreq = 3,
# Lambdas_Roughness = exp(-5:5),
# outlierTrimPct = 0.05,
# N_Bootstrap_for_Outlier = 500,
# RE_REGISTER = FALSE,
# Convergence_Threshold = 1e-05,
# SimMeanDiff_Threshold = 0.001,
# MinSimilarityThreshold = 0.25,
# MAX_ITERATION = 4
#)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.