Adding contaminated data to randomly produce conflicts

[igrangel]

Following this idea would be good if we could add a method that introduces some contaminated data by pairs, e.g., two attributes with same refSemantic and different names, etc.

[omarrana]

I have added this method for RoleClass attribute. It will change RoleClassAttribute Names in one of the seeds which give us the above scenario.
However, I am not sure about the GoldStandard here.

Please add this in config.ttl if you want to test.
aml:ContaminateRoleClassAttribute "1";

[igrangel]

Would be good if we can add the contaminated data to Attributes and Internal Elements since they are the most common AML elements.

[igrangel]

Would be possible to have the percents of the injected contaminated data we have?
For example, that we can inject 20% and 40%, etc, with respect to the total number of elements?

[omarrana]

Hi adding to other elements is not problem.
Currently it contaminates only generated data. So for role class I generated 2 attributes . then 2 are contaminated. We know the count of contamination through config

[igrangel]

Can we know the total number of elements?

[omarrana]

if we count the elements of seed.aml , then we know out of say 10 elements 2 were contaminated then we can . But i will have to run the results again since its overwritten now.

[omarrana]

close it please