search

anuzzolese / oke-challenge-2016

22 stars 8 forks source link

Task 2, type alignment : which dataset to use ? #10

Open Flutifioc opened 8 years ago

Flutifioc commented 8 years ago

Hello again,

After speaking of the previous issue (https://github.com/anuzzolese/oke-challenge-2016/issues/9) with my colleagues, we found out that the question is actually double.

In the README.md page, in the task 2 section, the link "DOLCE+DnS Ultra Lite classes" points to a d0.owl file, containing only a dozen of classes. However, when we look for "DOLCE+DnS Ultra Lite" on the internet, we find a DUL.owl file (http://www.ontologydesignpatterns.org/ont/dul/DUL.owl), whose label is "DOLCE+DnS Ultralite", but containing no d0 class.

The question is : what is the proper dataset to use for this subtask ? Until now, we used only the second one in our program (DUL.owl), but are we supposed to use also d0.owl, or only d0.owl, or only DUL.owl ? The gold standard contains both d0 and dul classes, so we are inclined to think that we should use both (as mentioned in the previous issue), could you confirm this ?

Thanks in advance !

anuzzolese commented 8 years ago

Dear @Flutifioc,

you are supposed to use both d0 and Dolce. However, d0 provides a generalisation over Dolce classes. The complete list of types with examples has been added in a table inside the description of task 2 (cf. https://github.com/anuzzolese/oke-challenge-2016#task-2)

Flutifioc commented 8 years ago

Thank you very much for both your answers.