Browsing by Subject "Spatial Big Data"
Now showing 1 - 2 of 2
- Results Per Page
- Sort Options
Item Geospatial Data Science to Identify Patterns of Evasion(2018-01) Eftelioglu, EmreOver the last decade, there has been a significant growth in the availability of cheap raw spatial data in the form of GPS trajectories, activity/event locations, temporally detailed road networks, satellite imagery, etc. These data are being collected, often around the clock, from location-aware applications, sensor technologies, etc. and represent an unprecedented opportunity to study our economic, social, and natural systems and their interactions. For example, finding hotspots (areas with unusually high concentration of activities/events) from activity/event locations plays a crucial role in epidemiology since it may help public health officials prevent further spread of an infectious disease. In order to extract useful information from these datasets, many geospatial data tools have been proposed in recent years. However, these tools are often used as a “black box”, where a trial-error strategy is used with multiple approaches from different scientific disciplines (e.g. statistics, mathematics and computer science) to find the best solution with little or no consideration of the actual phenomena being investigated. Hence, the results may be biased or some important information may be missed. To address this problem, we need geospatial data science with a stronger scientific foundation to understand the actual phenomena, develop reliable and trustworthy models and extract information through a scientific process. Thus, my thesis investigates a wide-lens perspective on geospatial data science, considering it as a transdisciplinary field comprising statistics, mathematics, and computer science. This approach aims to reduce the redundant work across disciplines as well as define scientific boundaries of geospatial data science to distinguish it from being a black box that claims to solve every possible geospatial problem. In my proposed approaches, I used ideas from those three disciplines, e.g. spatial scan statistics from statistical science to reduce chance patterns in the output and provide statistical robustness; mathematical definitions of geometric shapes of the patterns, which maintain correctness and completeness; and computational approaches (along with prune and refine framework and dynamic programming ideas) to scale up to large spatial datasets. In addition, the proposed approaches incorporate domain-specific geographic theories (e.g., routine activity theory in criminology) for applicability in those domains that are interested in specific patterns, which occur due to the actual phenomena, from geospatial datasets. The proposed techniques have been applied to real world disease and crime datasets and the evaluations confirmed that our techniques outperform current state-of-the-art such as density based clustering approaches as well as circular hotspot detection methods.Item Spatial Big Data Analytics: Classification Techniques for Earth Observation Imagery(2016-08) Jiang, ZheSpatial Big Data (SBD), e.g., earth observation imagery, GPS trajectories, temporally detailed road networks, etc., refers to geo-referenced data whose volume, velocity, and variety exceed the capability of current spatial computing platforms. SBD has the potential to transform our society. Vehicle GPS trajectories together with engine measurement data provide a new way to recommend environmentally friendly routes. Satellite and airborne earth observation imagery plays a crucial role in hurricane tracking, crop yield prediction, and global water management. The potential value of earth observation data is so significant that the White House recently declared that full utilization of this data is one of the nation's highest priorities. However, SBD poses significant challenges to current big data analytics. In addition to its huge dataset size (NASA collects petabytes of earth images every year), SBD exhibits four unique properties related to the nature of spatial data that must be accounted for in any data analysis. First, SBD exhibits spatial autocorrelation effects. In other words, we cannot assume that nearby samples are statistically independent. Current analytics techniques that ignore spatial autocorrelation often perform poorly such as low prediction accuracy and salt-and-pepper noise (i.e., pixels predicted as different from neighbors by mistake). Second, spatial interactions are not isotropic and vary across directions. Third, spatial dependency exists in multiple spatial scales. Finally, spatial big data exhibits heterogeneity, i.e., identical feature values may correspond to distinct class labels in different regions. Thus, learned predictive models may perform poorly in many local regions. My thesis investigates novel SBD analytics techniques to address some of these challenges. To date, I have been mostly focusing on the challenges of spatial autocorrelation and anisotropy via developing novel spatial classification models such as spatial decision trees for raster SBD (e.g., earth observation imagery). To scale up the proposed models, I developed efficient learning algorithms via computational pruning. The proposed techniques have been applied to real world remote sensing imagery for wetland mapping. I also had developed spatial ensemble learning framework to address the challenge of spatial heterogeneity, particularly the class ambiguity issues in geographical classification, i.e., samples with the same feature values belong to different classes in different spatial zones. Evaluations on three real world remote sensing datasets confirmed that proposed spatial ensemble learning outperforms current approaches such as bagging, boosting, and mixture of experts when class ambiguity exists.