Efficient Text Search over Temporally Versioned Document Collections
Ozlem Gur
MSc.Student
Computer Engineering Department
Bilkent University
Until recently, text search over temporally versioned document collections such as web archives has received little attention. Currently, most search engines do not allow time-range queries. If the user submits a valid or transaction time-range query, the information retrieval system still should be able to respond in a reasonable amount of time. Transaction time is the time interval from data is inserted to the database until it is deleted. Valid time is the time interval during which the information is valid in the real world. In the Web context, valid time can be thought as the modification time of a Web resource such as an HTML document and transaction time is the modification detection time of the crawler. Since obtaining transaction time is easier and more feasible, majority of the studies focus on transaction time queries. To be able to efficiently answer time-range queries, temporal-web inverted indices are proposed. In this project, we are willing to analyze the feasibility of temporal-web inverted-indices and investigate the alternative ways of implementing a temporal-web inverted-index so that query response time and space allocation requirements could be improved.
DATE:
3 December, 2007, Monday@ 16:15
PLACE:
EA 409