Author
Holzbauer, Herbert O.
Other Contributors
Szymanśki, Bolesław; Cutler, Barbara M.; Magdon-Ismail, Malik; Korniss, Gyorgy;
Date Issued
2016-12
Subject
Computer science
Degree
PhD;
Terms of Use
This electronic version is a licensed copy owned by Rensselaer Polytechnic Institute, Troy, NY. Copyright of original work retained by author.;
Abstract
A social network is a collection of users and relationships between them, typically viewed as a graph. The most common type of relationship is a friendship, such as seen in popular social networking platforms like Facebook. However, these networks exist in a variety of contexts both online and offline. Regardless of the medium or context, they allow us to quantify relationships between individual humans. These social networks, and the underlying communities they describe, contribute to our understanding of human behavior. Specifically we consider the impact of these ties when they are embedded in space.; We also use Gowalla to predict economic performance in the form of United States (US) Gross Domestic Product (GDP) using both the geographic location and social links of Gowalla users. We find that long ties (those that cross state boundaries) are an invaluable tool in estimation of GDP. We also discuss use of the predictors we develop in two other economic contexts, but ultimately find that these metrics are ill-suited for our approach and we explore why.; We then directly use a location-based service with a social network, Gowalla, towards two goals. The first is to analyze artificial social searches inspired by Milgram's small world experiments (delivering a package to a target using real acquaintances). By creating protocols which a rational agent could use for making forwarding decisions, we are able to explore the effect of several network features, and of partial knowledge of friends-of-friends on the social search.; We first demonstrate this indirectly in incentivized participatory sensing (where humans voluntarily perform sensing tasks in exchange for rewards) by leveraging human mobility, intelligence, and technology to select and collect evidence of events and phenomena occurring in the real world. We utilize a set of traces derived from actual human mobility and compare this to previously published work in which we had a similar system but employed random synthetic mobility. Based on the differences, we propose that human mobility is a critical part of understanding opportunistic networks such as the participatory sensing problem we studied.;
Description
December 2016; School of Science
Department
Dept. of Computer Science;
Publisher
Rensselaer Polytechnic Institute, Troy, NY
Relationships
Rensselaer Theses and Dissertations Online Collection;
Access
Restricted to current Rensselaer faculty, staff and students. Access inquiries may be directed to the Rensselaer Libraries.;