So it sounds like you want the SVD of the item-item similarity matrix? Sure, you can use Mahout for that. If you are not in Hadoop land then look at SVDRecomnender to crib some related code. It is decomposing the user item matrix though.
But for this special case of a symmetric matrix your singular vectors are the eigenvectors which you may find much easier to compute. I might restate the interpretation. The 'size' of these vectors is not what matters to your question. It is which elements (items) have the smallest vs largest values . On Jul 10, 2011 3:08 AM, "Lance Norskog" <[email protected]> wrote:
