We may need to update the algorithm so that the tests pass. Current tests that fail are:
[ ] test_source_to_source_disjoint
Currently includes the S1 to S2 edge in the final network.
[ ] test_weighted
Currently includes all the lowest weighted edges in the final product, should be only selecting the most confident network for a source-target pair (highest weight) when all other things are equal. I may not be understanding the weights properly though or getting confused with the confidence score.
Some ideas from BTB:
Confidence score calculation to determine the final network:
There are some "expected output" files for these test cases that I believe should match the "output" files for testing.
We may need to update the algorithm so that the tests pass. Current tests that fail are:
[ ] test_source_to_source_disjoint Currently includes the S1 to S2 edge in the final network.
[ ] test_weighted Currently includes all the lowest weighted edges in the final product, should be only selecting the most confident network for a source-target pair (highest weight) when all other things are equal. I may not be understanding the weights properly though or getting confused with the confidence score.
Some ideas from BTB: Confidence score calculation to determine the final network:
There are some "expected output" files for these test cases that I believe should match the "output" files for testing.