Additional functionality

Having got PROPheT installed and running, it is now time to explore the population functions of the tool, which enable the user to use external resources to add information to their ontology and which are some of the main contributions of the tool:

Since these functions may be more complicated, the demonstration provides both a quick access summary as well as a worked example of each function.

The first option is to Search by Instance label, which enables the user to type search into PROPheT, which provides a ‘label’ (e.g., Vincent van Gogh) of the specific instance they are looking for (or rather the information from external ontologies which relates to the instance they are looking for). Using some of the functions from the ‘Guided Tour’, the user can refine their search, the external sources queried, and the types of results that will be returned, as well as identifying which local ontology to populate. Once results have been returned, the user may wish to modify some of the search parameters they have specified (e.g., to generate more specific or more numerous results), which would then be followed by the selection and mapping steps to populate the local ontology model. This process is performed in detail using the label ‘Vincent van Gogh’ in the demonstration.

In addition to searching for specific instances, it is also possible to Search by Class, which allows the larger-scale populating of a user’s ontology on the basis of ‘class’ or type of instance. The process is similar to that of searching by instance label, but the user needs to know additional information about the exact type of class that they are searching for in the external model (the prefix and class name). For example, if the user wanted to find artworks in their external model, in the case of DBpedia the user would need to search for ‘dbo:Artwork’. As before, the user may wish to modify their search before selecting and mapping instances for the population of their model.

A search functionality of PROPheT which provides a more sophisticated method of populating a local model is to Search by existing Instance. This combines some of the specificity of ‘Search by instance label’ with the broader results obtained by ‘Search by class’. Here a user selects an instance from their local model (e.g., for an artist that they may have called Picasso, ‘my_ont:Picasso’). From there, the user is then able to perform an ‘rdfs:label’ search, which will display labels in the external ontology model which are associated with the search instance in the local model. By selecting the most relevant of these labels for the current purpose (e.g., ‘ModernPainters’, ‘SpanishSculptors’), the user is then able to perform a search of the external model to reveal other artists which share the same labels as existing instance (Picasso). Once again, the user retains control over the import of information from external sources and so they need to select and map external instances in order to populate their model.

The final functionality available to users of PROPheT is to Enrich an existing Instance. Similar to ‘Search by existing instance’, this allows users to search for an instance from their local model in an external ontology; however rather than identifying other, similar instances, this function allows users to find additional information relating to their instance and to use this to enrich their local model. Selecting and mapping of external information is once again performed by the user before inclusion in to their local model.