Electronic Theses and Dissertations





Date of Award


Document Type


Degree Name

Doctor of Philosophy


Computer Science

Committee Chair

Vasile Rus

Committee Member

Dipankar Dasgupta

Committee Member

King-Ip Lin

Committee Member

Lan Wang


Traditional information retrieval methodology is guided by document retrieval paradigm, where relevant documents are returned in response to user queries. This paradigm faces serious drawback if the desired result is not explicitly present in a single document. The problem becomes more obvious when a user tries to obtain complete information about a real world entity, such as person, company, location etc. In such cases, various facts about the target entity or concept need to be gathered from multiple document sources. In this work, we present a method to extract information about a target entity based on the concept retrieval paradigm that focuses on extracting and blending information related to a concept from multiple sources if necessary. The paradigm is built around a generic notion of concept which is defined as any item that can be thought of as a topic of interest. Concepts may correspond to any real world entity such as restaurant, person, city, organization, etc, or any abstract item such as news topic, event, theory, etc. Web is a heterogeneous collection of data in different forms such as facts, news, opinions etc. We propose different models for different forms of data, all of which work towards the same goal of concept centric retrieval. We motivate our work based on studies about current trends and demands for information seeking. The framework helps in understanding the intent of content, i.e. opinion versus fact. Our work has been conducted on free text data in English. Nevertheless, our framework can be easily transferred to other languages.


Data is provided by the student.

Library Comment

dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.