Mining sensor datasets with spatiotemporal neighborhoods

Michael Patrick Mcguire,Vandana Janeja,Aryya Gangopadhyay

doi:10.5311/josis.2013.6.94

Abstract

Many spatiotemporal data mining methods are dependent on how relationships between a spatiotemporal unit and its neighbors are defined. These relationships are often termed the neighborhood of a spatiotemporal object. The focus of this paper is the discovery of spatiotemporal neighborhoods to find automatically spatiotemporal sub-regions in a sensor dataset. This research is motivated by the need to characterize large sensor datasets like those found in oceanographic and meteorological research. The approach presented in this paper finds spatiotemporal neighborhoods in sensor datasets by combining an agglomerative method to create temporal intervals and a graph-based method to find spatial neighborhoods within each temporal interval. These methods were tested on real-world datasets including (a) sea surface temperature data from the Tropical Atmospheric Ocean Project (TAO) array in the Equatorial Pacific Ocean and (b) NEXRAD precipitation data from the Hydro-NEXRAD system. The results were evaluated based on known patterns of the phenomenon being measured. Furthermore, the results were quantified by performing hypothesis testing to establish the statistical significance using Monte Carlo simulations. The approach was also compared with existing approaches using validation metrics namely spatial autocorrelation and temporal interval dissimilarity. The results of these experiments show that our approach indeed identifies highly refined spatiotemporal neighborhoods.

Highlights

Spatiotemporal data from automatic sensing devices has become prevalent in many domains such as climatology, hydrology, transportation planning, and environmental science and we are currently experiencing a deluge of spatiotemporal data collected at increasingly numerous locations and fine temporal granularities
This paper presents a novel method to identify spatiotemporal intervals and neighborhoods
The approach first discovers temporal intervals based on spatial relationships; and second discovers spatial neighborhoods within the temporal intervals

Summary

Introduction

Spatiotemporal data from automatic sensing devices has become prevalent in many domains such as climatology, hydrology, transportation planning, and environmental science and we are currently experiencing a deluge of spatiotemporal data collected at increasingly numerous locations and fine temporal granularities. In the context of this paper, a sensor is defined as a device that automatically measures a physical quantity over time at a location. Finding spatiotemporal patterns in large sensor datasets helps identify physical processes governing the phenomenon being measured. As the data grows over time, spatiotemporal patterns become more difficult to analyze. Another challenge in spatiotemporal data mining is to identify the proper method for neighborhood generation. This paper focuses on the discovery of spatiotemporal neighborhoods where, in space, the neighborhood is generally a set of locations that are proximal and have similar characteristics. Our notion of a spatiotemporal neighborhood is distinct from the traditional notions since we consider both a spatial characterization as well as a temporal characterization of neighborhoods

Objectives

Methods

Results

Discussion

Conclusion