Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form


Sharker, Md Monir H (2018) AN ADAPTIVE FRAMEWORK FOR REAL-TIME SPATIOTEMPORAL BIG DATA ANALYTICS. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Download (4MB) | Preview


Due to advancements in and widespread usage of technologies such as smartphones, satellites, smart sensors, and social networks, collection of spatiotemporal data is growing rapidly. Such massive spatiotemporal data require appropriate techniques and technologies for their efficient analysis and processing. Analyzing massive spatiotemporal data efficiently and effectively is challenging since the data changes dynamically over space and time whereas, often, decisions followed by the analysis need to be made under real-time constraints. Compared to non-spatial data, spatiotemporal data, among other unique characteristics, are multidimensional (x, y, attributes, time) in nature, complex in structures and behaviors, and provides details at different resolutions and scales. These characteristics together make analyzing and processing massive spatiotemporal data in real time a challenging task. Resorting to high-performance computing (HPC) is a common approach for handling this computing challenge but to determine optimal solutions through data and computation analysis, appropriate analytics and computing solutions are needed.
In this dissertation, we proposed a framework which is basically a platform providing spatiotemporal data-intensive analytics for data- and compute-intensive applications that require computation under real-time constraints on given computing resources. The framework is a layered structure consisting of four interrelated components (layers); three on analytics and one on adaptive computing. A graph-based approach is developed as the foundation of the analytics components which are: efficient analytics – providing acceptable solutions based on current data in the absence of historical data; predictive analytics – providing near-optimal solutions by learning from the patterns of historical data and predicting based on the learning; meta-analytics – providing optimal solutions by analyzing pattern of past data patterns; and adaptive computing that ensures appropriate analytics are applied and computation is completed in real time on available computing resources.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Sharker, Md Monir Hmhs37@pitt.edumhs370000-0001-9198-601X
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairKarimi, Hassan
Committee MemberMelhem, Rami
Committee MemberDruzdzel, Marek
Committee MemberMunro, Paul
Date: 24 January 2018
Date Type: Publication
Defense Date: 14 August 2017
Approval Date: 24 January 2018
Submission Date: 13 December 2017
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 158
Institution: University of Pittsburgh
Schools and Programs: School of Computing and Information > Information Science
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Adaptive Framework, Spatiotemporal, Big Data Analytics, Real-Time Computing, Predictive Analytics, Machine Learning
Date Deposited: 24 Jan 2018 16:27
Last Modified: 24 Jan 2018 16:27


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item