NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED527468
Record Type: Non-Journal
Publication Date: 2008
Pages: 169
Abstractor: As Provided
Reference Count: 0
ISBN: ISBN-978-1-1096-6058-6
ISSN: N/A
Large Scale IR Evaluation
Pavlu, Virgil
ProQuest LLC, Ph.D. Dissertation, Northeastern University
Today, search engines are embedded into all aspects of digital world: in addition to Internet search, all operating systems have integrated search engines that respond even as you type, even over the network, even on cell phones; therefore the importance of their efficacy and efficiency cannot be overstated. There are many open possibilities for new ideas, implementations and optimizations in this area; that, combined with the immense interest from both academia and corporations, makes it a very attractive research field. My PhD work is focused on search quality. I have developed algorithms and models for efficient evaluation, estimation of query difficulty, metasearch, and exploration of relevant patterns; I also participated to collection-building efforts for research purposes. This thesis presents a unified view of many of these tasks, often making use of intrinsic connections between them. We propose solutions that make current technologies scalable with data growth, a certainty of the close future. Some of these solutions have been already implemented and used by the research community; certain methods made possible research endeavors that would have been otherwise unworkable. The main result of this thesis is the proposal of two methods for pooling and evaluation of IR systems. Both approaches are well studied in other domains, and have a good mathematical foundation. Both offer a strong direction for efficient evaluation, without which IR experimentation cannot be performed on a large scale. An additional benefit of the sampling methods is that estimated variance leads to a confidence in the evaluation, a first within all IR evaluation technology. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]
ProQuest LLC. 789 East Eisenhower Parkway, P.O. Box 1346, Ann Arbor, MI 48106. Tel: 800-521-0600; Web site: http://www.proquest.com/en-US/products/dissertations/individuals.shtml
Publication Type: Dissertations/Theses - Doctoral Dissertations
Education Level: Higher Education
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A