I am looking for bicliques of particular orders, eg 4 x 4 or 7 x 5, and would be happy to know that there is such a biclique except for one or two missing edges.
Knowing that there is a 12 x 12 biclique in the data except for 3 missing edges would be significant.
Would you be able to extend your algorithm to this problem?
By the way, I could look at each combination of left vertices taken 7 at a time, and each combination of right vertices taken 5 at a time, and return any pair of combinations for which the number of edges is at least 30 (say) of 35. But I'd prefer your fast algorithm do it.
As a minor matter, your arguments include left_min and right_min (actually _least), and I can easily test for specific values of left and right vertex counts. But might you also provide arguments left_max and right_max to narrow the search for bicliques and approximate bicliques?
Hi Yuping,
I am looking for bicliques of particular orders, eg 4 x 4 or 7 x 5, and would be happy to know that there is such a biclique except for one or two missing edges.
Knowing that there is a 12 x 12 biclique in the data except for 3 missing edges would be significant.
I think I can see how to modify your Algorithm 1 in https://bmcresnotes.biomedcentral.com/articles/10.1186/s13104-020-04955-0 to identify these approximate bicliques. Is that something you have considered?
Would you be able to extend your algorithm to this problem?
By the way, I could look at each combination of left vertices taken 7 at a time, and each combination of right vertices taken 5 at a time, and return any pair of combinations for which the number of edges is at least 30 (say) of 35. But I'd prefer your fast algorithm do it.
As a minor matter, your arguments include left_min and right_min (actually _least), and I can easily test for specific values of left and right vertex counts. But might you also provide arguments left_max and right_max to narrow the search for bicliques and approximate bicliques?
Thanks again, Colin