Implements world min-hash sparsification for input matches. For very large all-to-all alignments, we can randomly remove most matches and still have a coherent graph. By sparsifying at the level of exact matches rather than mappings, we retain long range information from the input alignment set.
Implements world min-hash sparsification for input matches. For very large all-to-all alignments, we can randomly remove most matches and still have a coherent graph. By sparsifying at the level of exact matches rather than mappings, we retain long range information from the input alignment set.