Sparse rank matrix factoristation (Sparse RMF) decomposes a rank matrix into a product of two smaller matrices, one of which contains k partial rankings of columns. The new representation reveals the structure of the data and summarises the original matrix in a succinct way.
Sparse RMF is formalised as an optimisation problem. It is solved by a greedy algorithm, which uses Integer Linear Programming technique.
Sourcecode
Contact
thanh dot levan at cs dot kuleuven dot be