GeoAI and Deep Learning Symposium: Spatial-Temporal Modeling and Data Mining II

Type: Paper
Theme:
Sponsor Groups: Spatial Analysis and Modeling Specialty Group, Geographic Information Science and Systems Specialty Group
Poster #:
Day: 4/4/2019
Start / End Time: 5:00 PM / 6:40 PM
Room: Capitol Room, Omni, East
Organizers: Diep Dao, Daniel Runfola
Chairs: Diep Dao

Description

Geographic information science has been experiencing an unprecedented increase in data accessibility and computational capability. This enables us to apply and enhance various data mining techniques such as rule learning, pattern recognition, classification, similarity and distances, and clustering, for spatial and spatio-temporal data analytics and knowledge discovery. However, unique characteristics of spatial data, including spatial and spatio-temporal relations, correlations, heterogeneity, hierarchy, interactions, and mix of data at different formats and resolutions, prevent a direct deployment of traditional data mining approaches in spatial and spatio-temporal mining problems. Spatial and spatio-temporal data mining on big datasets, in addition, requires implementation under high performance computing environment with unique challenges for data preparation, data manipulation, data simulation, large-scale and multi-dimensional vector analysis, and memory allocation, to name a few.

This session welcomes contributions exploring theoretical and practical issues of data mining approaches for spatial or spatio-temporal datasets. Applications within a broad range of GIScience applied domains in support of society and/or environmental challenges are welcomed. We hope to create an opportunity for researchers to exchange ideas around the use of data mining approach as an inductive approach toward big geographical data analysis.

Specific topics include (but are not limited to):
- Applications of data mining approaches for GIScience big data
- Handling spatial and spatio-temporal dimensions in data mining
- Developments of novel spatial or spatio-temporal data mining techniques
- Understand text-based mined results for evaluation through visualization
- From data mining to knowledge discovery: start the validation process with support from domain knowledge
- Applications of High Performance Computing (HPC) tools
- Development of tools or approaches to implementing spatial data mining in HPC environments
- Computational challenges with large vector data
- Computational challenges with global-scope, high resolution analyses of imagery data


Agenda

Type Details Minutes Start Time
Presenter Yuchuan Huang*, University of Minnesota - Minneapolis, Ying Song, University of Minnesota - Minneapolis, Investigating Relationships between Bike-sharing and Public Transit: A Spatial-Temporal Approach 20 5:00 PM
Presenter Diep Dao*, University of Colorado - Colorado Springs, Jean-Claude Thill, University of North Carolina - Charlotte, Mining for Spatial Associations in Urban Residential Auto Theft Analysis 20 5:20 PM
Presenter Fernando Sanchez-Trigueros*, University of Arizona - Geography & Development, Evolving a genetic algorithm to solve a multiple knapsack optimization problem with allocation frictions, and its application to maximizing spatial accessibility to services with minimal reallocation costs 20 5:40 PM
Presenter Daniel Runfola*, The College of William and Mary | AidData, Geeta Batra, Global Environment Facility, Anupam Anand, Global Environment Facility, Causal Trees: Exploring Causal Linkages using Machine Learning and Spatial Data 20 6:00 PM
Presenter Seth Goodman*, College of William and Mary, Daniel Runfola, College of William and Mary, A Comparison of Landsat 7 Gap Filling Methods for Convolutional Neural Network Applications 20 6:20 PM

To access contact information login