Abstract
In this paper we study the identifiability of the Paralind model with sparse interaction matrices (i.e. S-Paralind). We provide some theoretical results on how to obtain the sparsest interaction matrices in some particular configurations and when these matrices are unique. These results could be use for the design and analysis of l0-based decomposition algorithms.
Cite
CITATION STYLE
Miron, S., & Brie, D. (2015). Some rank conditions for the identifiability of the sparse paralind model. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9237, pp. 41–48). Springer Verlag. https://doi.org/10.1007/978-3-319-22482-4_5
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
