Skip to content

2.2 DomainIdentifier

Kannan Ramamoorthy edited this page Jan 10, 2019 · 4 revisions


The purpose of the domain identifier interface is to encapsulate classes that aim to detect a domain whose pages might contain information pertaining to a particular set of positive examples, of which each is a pair of resources that stand in a user-defined relation (e.g., actors and the movies they played in). The exact definition of the interface can be foud here.

Parameters (no optional parameters!)

Every domain identifier requires the following parameters:

  1. A property p: This property is that for which REX is supposed to extract more instance knowledge.
  2. A set of positive examples: This set of pairs of resources contais the subject and object that are to be used to detect a domain that will allow extraction more instance knowledge for p.
  3. A set of negative examples: This set is optional and contains resources that should be not extracted.
  4. A boolean that switch the use of a cache on and off.

The correspoding signature corresponds to the parameters described above:

       URL getDomain(Property p, Set<Pair<Resource, Resource>> posExamples, Set<Pair<Resource, Resource>> negExamples, boolean useCache);

We provide an automatic implementation of the domain identifier which is based on the Google API and can be found here. More implementations are welcome.

You can’t perform that action at this time.