Searching the Semantic Web

Download Report

Transcript Searching the Semantic Web

Searching
the
Semantic Web
Introduction
 Research Focuses:


IE
Ontologies (creating, languages, merging,
storing, querying)
 Next Sep: Using the Semantic Web


How will we use <annotated> pages for
useful tasks?
How will we build Semantic Web Services?
Introduction
 Example:
Traveler wants to know:

Cruises in San Francisco Bay
 Company
Name
 Price
 Location
 Schedule

How will a Searching Service provide that
information?
Thesis Statement (revised for ppt)
 Take Next Logical Step In Semantic Web
 Build Searching Service


Use Google as search engine
Assume:
 Most
(if not all) pages are annotated with OWL
 Primary Objects are known
 Service Has a Global Ontology
Outline
 1. Semantic Querying
 2. Crawling
 3. Indexing
 4. Returning Results
1. Semantic Querying
 How do you know what the user meant?
How do you distinguish the subject?

Extremes: SQL queries

Accept anything
 RDF = (Subject, Predicate, Object)
 User provides desired subject, and objects
(attributes)
2. Crawling
 Google’s Googlebot does real crawling
 We look at results from normal Google
search
 How to find matches on annotated pages?
2. Crawling
3. Indexing
 Index on Records Given from Crawling
 Ideas:


Fully Filled Records First
Closeness of Concept Match
 Measure
by distance from referenced
ontology mapping to user words mapping in
Global Ontology
4. Returning Results
 Returns Dynamic Page of Records:
Name: Angel Island Tiburon-Ferry
Price: $8.00
Location:
Schedule:
21 Main Street
Hourly
http://www.sourcepage.com
 Records may be ordered according to chosen
attribute
Contributions
 Enhance the ability to access information
on the Web
 Step closer to machine understanding
 Building block for Semantic Web Services
 Past research put to use