Study of result presentation and interaction for aggregated search

Sushmita, Shanu (2012) Study of result presentation and interaction for aggregated search. PhD thesis, University of Glasgow.

Abstract

The World Wide Web has always attracted researchers and commercial search engine companies due to the enormous amount of information available on it. "Searching" on web has become an integral part of today's world, and many people rely on it when looking for information. The amount and the diversity of information available on the Web has also increased dramatically. Due to which, the researchers and the search engine companies are making constant efforts in order to make this information accessible to the people effectively.

Not only there is an increase in the amount and diversity of information available online, users are now often seeking information on broader topics. Users seeking information on broad topics, gather information from various information sources (e.g, image, video, news, blog, etc). For such information requests, not only web results but results from different document genre and multimedia contents are also becoming relevant. For instance, users' looking for information on "Glasgow" might be interested in web results about Glasgow, Map of Glasgow, Images of Glasgow, News of Glasgow, and so on.

Aggregated search aims to provide access to this diverse information in a unified manner by aggregating results from different information sources on a single result page. Hence making information gathering process easier for broad topics.

This thesis aims to explore the aggregated search from the users' perspective. The thesis first and foremost focuses on understanding and describing the phenomena related to the users' search process in the context of the aggregated search. The goal is to participate in building theories and in understanding constraints, as well as providing insights into the interface design space. In building this understanding, the thesis focuses on the click-behavior, information need, source relevance, dynamics of search intents. The understanding comes partly from conducting users studies and, from analyzing search engine log data.

While the thematic (or topical) relevance of documents is important, this thesis argues that the "source type" (source-orientation) may also be an important dimension in the relevance space for investigating in aggregated search. Therefore, relevance is multi-dimensional (topical and source-orientated) within the context of aggregated search. Results from the study suggest that the effect of the source-orientation was a significant factor in an aggregated search scenario. Hence adds another dimension to the relevance space within the aggregated search scenario.
The thesis further presents an effective method which combines rule base and machine learning techniques to identify source-orientation behind a user query.

Furthermore, after analyzing log-data from a search engine company and conducting user study experiments, several design issues that may arise with respect to the aggregated search interface are identified. In order to address these issues, suitable design guidelines that can be beneficial from the interface perspective are also suggested.

To conclude, aim of this thesis is to explore the emerging aggregated search from users' perspective, since it is a very important for front-end technologies. An additional goal is to provide empirical evidence for influence of aggregated search on users searching behavior, and identify some of the key challenges of aggregated search. During this work several aspects of aggregated search will be uncovered. Furthermore, this thesis will provide a foundations for future research in aggregated search and will highlight the potential research directions.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Aggregated search, relevance, source-orientation, user interaction, search interface, information retrieval.
• Subjects: T Technology > T Technology (General); Q Science > Q Science (General)
• Colleges/Schools: College of Science and Engineering > School of Computing Science
• Supervisor's Name: Mounia, Professor Lalmas and Joemon, Professor M. Jose
• Date of Award: 2012
• Depositing User: Ms Shanu Sushmita
• Unique ID: glathesis:2012-3289
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 29 Mar 2012
• Last Modified: 10 Dec 2012 14:05
• URI: https://theses.gla.ac.uk/id/eprint/3289