In order to join virtual sessions, you must be registered and logged-in(Were you registered for the in-person meeting in Denver? if yes, just log in.) 
Note: All session times are in Mountain Daylight Time.

Finding anomalies based upon multi-scale spatial dependence

Authors: Boleslo E. Romero*, University of New Mexico
Topics: Geographic Information Science and Systems, Spatial Analysis & Modeling
Keywords: GeoAI, anomaly, spatial outlier, multiscale, spatial dependence
Session Type: Paper
Presentation File: No File Uploaded


Spatial outliers are local extrema, or anomalies. Their representation in spatial data can be complicated by the effects of scale. If the scale of analysis is broader than an outlier it can be masked and not well represented. Conversely, if the outlier is broader than the scale of analysis its representation swamps across numerous data, which might limit the performance of spatial outlier detection methods.

In the latter case, regions near the peak of the outlier might exhibit higher degrees of spatial autocorrelation compared to other parts of the outlier, such as the side. If so, there is a potential for identifiable parts of outliers to support a process of locating outliers that span across regions represented by numerous data.

The assertion that identifiable parts of outliers exist is explored by testing the hypothesis that discernible patterns of spatial autocorrelation do not exist in portions of outliers. A metric of multi-scale spatial dependence is presented and employed in an experiment involving simulated outliers. The results suggest that identifiable parts of outliers are detectable in situations where the outlier spans across a region represented by numerous data.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login