PCA on time series

[battlestar]

Hi, excuse my poor english...
I need to reduce number of indicators (and to perform a PCA) on dataset:
year ind(1) ind(2) ... ind(n)
2000 x(1,1) x(1,2) ... x(1,n)
.....
2010 x(11,1) x(11,2) ... x(11,n)
x(i,j) are quantitative values about indicators like production, resources, ecc.
How year values must be evaluated? I use FactoMineR and I'd like to know if "year" must be considered a quantitative variable or a supplementary quant. variable, or.... ?
Thanks for Your attention

 

[terzi]

Hi battlestar,

I don't know much about it, but you may want to look into Functional Principal Components Analysis:

http://edoc.hu-berlin.de/master/benko-michael-2004-12-20/PDF/benko.pdf

 

[battlestar]

Hi, thanks. Anyway I've found a work in which the plot of scores for the principal components 1 and 2 seems to confirm that years must be used as labels/qualitative variables (in attachment).