PhD Qualifying Examination


"Data Integration in the Hidden Web"

Miss Qiong Huang


Abstract:

Accessing information in the hidden Web requires not only search for
relevant data but also integration of heterogeneous data sources.
Compared with the traditional data integration scenario, two major issues
of which are schema matching and entity resolution, web data integration
adds the following new challenges: source modeling and selection, source
querying and data extraction. To prepare for domain-based data
integration, source selection first classifies similar data sources while
source modeling extracts the schemas from the various data source
interfaces. For schema matching, many holistic methods take into account
the scalability problem specific to hidden Web data integration. Next,
source querying automatically simulates user inputs to retrieve relevant
data results. Once the results are returned in web pages, data extraction
techniques are employed for collecting data entities. Finally, data
entities from different data sources are matched and merged together. In
this survey, we first study the basic characteristics of hidden web data
integration and then review state-of-the-art techniques for each of these
sub problems. These techniques focus on the modeling of the corresponding
sub-problem or on how to identify and exploit various kinds of
information. We also point out some open research issues.


Date:     		Thursday, 18 January 2007

Time:                   2:00p.m.-4:00p.m.

Venue:                  Room 3501
			lifts 25-26

Committee Members:      Prof. Frederick Lochovsky (Supervisor)
			Dr. Lei Chen (Chairperson)
			Prof. Dik Lun Lee
			Dr. Qiong Luo


**** ALL are Welcome ****