Dear all, I observe for each observation several joint distributions of two multinomial random variables (5x5 matrices). Right now, data are arranged so that I have 20 columns for each joint distribution by observation, which is not practical. I would like to work with matrices that would be indexed by observation, just as I work with vectors and matrices for regression like analysis. How should I proceed? With a list of list of matrices and how to build it in a systematic way (n=3000 and I observe up to 10 joint distributions by observation)? I'm not familiar with such procedure and I don't know where I should start reading. Thanks! Best, SL
SL
2011-Jan-12 14:57 UTC
[R] Fwd: vector or list of matrices corresponding to an observation
You're right Jim. I have now started to work with list of matrices. I'm not sure that it is a clean and nice code but it works. I was wondering if there was a synthetic but advanced tutorial on list() and associated functions? Stephane ---------- Forwarded message ---------- From: jim holtman <jholtman at gmail.com> Date: 2011/1/12 Subject: Re: [R] vector or list of matrices corresponding to an observation To: SL <sl465 at yahoo.fr> An actual example of your data would be useful along with how you might want to access it. ?You can create a matrix of 'list' objects and these could contain your matrices, but not knowing what the data looks like, or how you intend to use it, make it hard to provide a solution.