[R] Dimensionality reduction with ProDenICA

[Rolf Turner]

On 18/01/17 11:17, Neverstop . wrote:
> Hello,
>
> I have a dataset with many variables and I'd like to do dimensionality
> reduction with Independent Component Analysis. There are many
> statistical methods to estimate the latent variables of the ICA model.
> I'm trying the R package "ProDenICA" that implements the penalized
> maximum likelihood method proposed by Hastie, Tibshirani and Friedman in
> Section 14.7.4 of the book "Elements of Statistical Learning". The
> documentation of the ProDenICA function says that the argument "k" is
> the "Number of components required, less than or equal to the number of
> columns of x". If I choose a value of k less than the number of colomns
> of x, I get an error message. It seems to me that I'm not using the
> function proDenICA() as it is meant to be used. Am I missing something?
>
> I've reproduced the problem with a smaller dataset here:
>
>  > library(MASS)
>  > data(crabs)
>  > str(crabs)
> 'data.frame':    200 obs. of  8 variables:
>   $ sp   : Factor w/ 2 levels "B","O": 1 1 1 1 1 1 1 1 1 1 ...
>   $ sex  : Factor w/ 2 levels "F","M": 2 2 2 2 2 2 2 2 2 2 ...
>   $ index: int  1 2 3 4 5 6 7 8 9 10 ...
>   $ FL   : num  8.1 8.8 9.2 9.6 9.8 10.8 11.1 11.6 11.8 11.8 ...
>   $ RW   : num  6.7 7.7 7.8 7.9 8 9 9.9 9.1 9.6 10.5 ...
>   $ CL   : num  16.1 18.1 19 20.1 20.3 23 23.8 24.5 24.2 25.2 ...
>   $ CW   : num  19 20.8 22.4 23.1 23 26.5 27.1 28.4 27.8 29.3 ...
>   $ BD   : num  7 7.4 7.7 8.2 8.2 9.8 9.8 10.4 9.7 10.3 ...
>  > X=crabs[,4:8]
>  > X=as.matrix(X)
>  > library(ProDenICA)
>  > out.proDen = ProDenICA(X, k = 2, whiten = TRUE, maxit = 20, trace=T)
> Error in solve.default(V, W) : 'a' (5 x 2) must be square
>
> I get the error with k = 1,2,3,4. The function works with k=5.

I have no idea how to help you --- sorry!!! --- but I would like to 
compliment you on the clarity of your extremely well structured 
question, complete with reproducible example.

My gut feeling is that there must be something wrong with the 
ProDenICA() function, but such is my ignorance that this gut feeling is 
not worth the paper it's written on! :-)

cheers,

Rolf Turner

-- 
Technical Editor ANZJS
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276