The function performs Non-Gaussian Component Analysis as described in Blanchard et.al. (2005).
ngca(data, L=c(1000,1000,1000), T=10, m=3, eps=1.5, npca=min(dim(x)[2],
dim(x)[1]), filter.time="None", filter.space=FALSE, method="temporal",
dg.trend = 2, h.space=3, h.time=3, keepv=TRUE, delta = NULL)Observation matrix (dimension Nxd)
Number basis functions in each of four classes.
Number of Fast ICA iterations
Number of non-Gaussian components.
Threshold (defaults to 1.5)
Reduce space to npca principal components. This can be
used to avoid standardizing by numerically singular covariance matrices.
In fMRI this allows to reduce the dimensionality assuming that
the interesting non-Gaussian directions are also characterized by
larger variances.
Choice of temporal filtering before analysis: "None", "Low", "Both", "High" (default "None")
Choice of spatial filtering before analysis: logical, default FALSE
Either "spatial" or "temporal". Specifies the type of NGCA to perform.
not yet documented
bandwidth for spatial filtering. default 3
bandwidth for temporal filtering. default 3
if TRUE intermediate results from fast ICA step are kept.
not yet documented
The function returns a list with components
Matrix containing the first m NGCA directions as columns.
Standard deviations of the principal components of the thresholded ICA directions
first m components of the rotated data
If keepv==TRUE the set of directions v^{(k)}
If keepv==TRUE the norm of each v^{(k)}.
The function performs Non-Gaussian Component Analysis as described in Blanchard et.al. (2006). The procedure uses four classes of basis functions, i.e. Gauss-Power3, Hyperbolic Tangent and the real and complex part of the Fourier class. See Blanchard et.al. (2005) for details.
Blanchard, G., Kawanabe, M., Sugiyama, M., Spokoiny, V. and M\"uller K.-R. (2005). In Search of Non-Gaussian Components of a High-Dimensional Distribution. Journal of Machine Learning Research. pp. 1-48.