Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

Author ORCID Identifier

N/A

AccessType

Open Access Dissertation

Document Type

dissertation

Degree Name

Doctor of Philosophy (PhD)

Degree Program

Electrical and Computer Engineering

Year Degree Awarded

2015

Month Degree Awarded

September

First Advisor

Lixin Gao

Subject Categories

Other Computer Engineering

Abstract

With the advances of technology and the popularity of the Internet, a large amount of data is being generated and collected. Much of these data is relational data, which describe how people and things, or entities, are related to one another. For example, data from sale transactions on e-commerce websites tell us which customers buy or view which products. Analyzing the known relationships from relational data can help us to discover knowledge that can benefit businesses, organizations, and our lives. For instance, learning the products that are commonly bought together allows businesses to recommend products to customers and increase their sales. Hidden or new relationships can also be inferred based on relational data. In addition, based on the connections among the entities, we can approximate the level of relatedness between two entities, even though their relationship may be hard to observe or quantify. This research aims to explore novel applications of relational data that will help to improve our life in various aspects, such as improving business operations, improving experiences in using online services, and improving health care services. In applying relational data in any domain, there are two common challenges. First, the size of the data can be massive, but many applications require that results are obtained within a short time. Second, relational data are often noisy and incomplete. Many relationships are extracted automatically from text resources, and hence they are prone to errors. Our goal is not only to propose novel applications of relational data but also to develop techniques and algorithms that will facilitate and make such applications practical. This work addresses three novel applications of relational data. The first application is to use relational data to improve user experiences in online video sharing services. Second, we propose the use of relational data to find entities that are closely related to one another. Such problems arise in various domains, such as product recommendation and query suggestion. Third, we propose the use of relational data to assist medical practitioners in drug prescription. For these applications, we introduce several techniques and algorithms to address the aforementioned challenges in using relational data. Our approaches are evaluated extensively to demonstrate their effectiveness. The approaches proposed in this work not only can be used in the specific applications we discuss but also can help to facilitate and promote the use of relational data in other application domains.

DOI

https://doi.org/10.7275/7370600.0

Share

COinS