Dissertation/Thesis Abstract

EventShop: An Information Engine for Processing Massive Heterogeneous Spatio-Temporal Data Streams
by Gao, Mingyan, Ph.D., University of California, Irvine, 2012, 202; 3544084
Abstract (Summary)

With the proliferation of diverse mobile devices and development of platforms like the Web and social networks, the data world is clearly moving away from its original underpinning in cyberspace and merging into the physical space. Enormous amounts of data pertaining to different characteristics across space and time are being generated and shared. These heterogeneous spatio-temporal data streams can be used to detect real-world situations and help people in taking appropriate actions in time for saving lives and resources. To allow such applications, a generic information engine that is able to process massive heterogeneous spatio-temporal data streams could be very useful. Building such an engine requires creating effective data models to represent spatio-temporal data streams and designing convenient operators and an efficient query system to combine and process a massive number of streams.

In this dissertation, we propose an information engine, named EventShop, for handling massive heterogeneous spatio-temporal data streams. First, we present the E-mage algebra, which consists of models for effectively representing spatial data and a set of expressive operators for processing streams, both based on a data representation called E-mage. Based on this algebra, we introduce the general system architecture and detailed component design in EventShop. In the context of the query system, we describe a cost-based multi-query optimization algorithm which allows discovery and sharing of re-usable computations and data sources among queries, based on computation and memory costs. We present the prototype system and demonstrate real-life and simulated applications, and we evaluate the efficiency of proposed multi-query optimization algorithms through extensive experiments. These applications and experiments show the efficacy of our work in processing spatio-temporal data streams about the physical world to satisfy various real needs in human life.

Indexing (document details)
Advisor: Jain, Ramesh
Commitee: Carey, Michael, Mehrotra, Sharad
School: University of California, Irvine
Department: Information and Computer Science - Ph.D.
School Location: United States -- California
Source: DAI-B 74/03(E), Dissertation Abstracts International
Subjects: Computer science
Keywords: Data streams, Situation recognition, Spatio-temporal analysis
Publication Number: 3544084
ISBN: 978-1-267-73715-1
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy