The purpose of semantic annotation is to add formally defined semantics into existing web documents so that machine agents can understand the content of web pages. The semantic web expects annotations to be connected with an ontology, which provides the semantics. Ontologies are commonly shared, explicitly defined, generic conceptualizations.
Our semantic annotation prototype system is based on Ontos, the BYU data-extraction engine, which is resilient to changes in the layout of web pages and immediately works with new pages in the same domain. This basic demo shows the capabilities of our semantic annotation tool.
To run the demo, users must choose a web page and a domain ontology, and then ask the system to do the annotation. During the process, a small window will pop up and show the progress of the annotation process. After the processing is is finished, users can view the annotated web page. The demo shows an annotated web page in two frames. The upper frame contains the web page with semantic annotation inside it. When the mouse is moved over an annotated data item on the page, the data item is highlighted and details about its connection to the ontology appear. The lower frame shows a table that contains all the recognized data instances within the annotated web page along with their interrelationships and their corresponding semantic categories. In addition, all the annotation information is stored inside an annotation ontology, represented in OWL. Therefore, machine agents can read the populated OWL ongology associated with the page rather than the page itself with embedded annotations, which is likely to be an easier task. In either case, the semantic annotation empowers them with the ability to understand the meaning of a web page.
A Windows version of our semantic annotator can be dowloaded here.