pick_dim: Number Of Principal Component Selection Based On Variation

In fChange: Change Point Analysis in Functional Data

Description

This function selects number of principal components based on the total variation explained (TVE). The functional data is projected onto a smaller number of principal curves via functional Principal Component Analysis (fPCA). This function picks the dimension of the projection space based on the desired total variation explained.

Usage

pick_dim(fdobj, TVE)

Arguments

fdobj	Functional data object, class of "fd"
TVE	Total Variation Explained. It must be in [0,1].

Details

This function is used to determine the dimension of the space that the functional data is projected onto based on the variation. One of the common treatments of the functional data is to transform it to multivariate objects using so called score vectors and utilize the multivariate techniques. This function will enable users to pick the dimension of the score vectors.

Value

d	Minimum number of Principle components needed in order to reach desired TVE
TVEs	Vector of TVEs. This has the same length of number of basis that the functional data is represented

References

Ramsay, James O., and Silverman, Bernard W. (2006), Functional Data Analysis, 2nd ed., Springer, New York

See Also

pca.fd

Examples

fdata1 = fun_IID(n=100, nbasis=21)
pick_dim(fdata1, 0.95)
fdata2 = fun_IID(n=100, nbasis=21, Sigma=3^-(1:21))
pick_dim(fdata2, 0.95)

fChange documentation built on May 2, 2019, 6:43 a.m.