Geographically Robust Hotspot Detection: A Summary of Results

Emre Eftelioglu, Xun Tang, Shashi Shekhar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

13 Scopus citations


Given a set of points in two dimensional space, a minimum radius, a minimum log likelihood ratio and a significance threshold, Geographically Robust Hotspot Detection (GRHD) finds hotspot areas where the concentration of points inside is significantly high. The GRHD problem is societally important for many applications including environmental criminology, epidemiology, etc. GRHD is computationally challenging due to the difficulty of enumerating all possible candidate hotspots and the lack of monotonicity property for the interest measure, namely the log likelihood ratio test. Related work may miss hotspots when hotspots are divided by geographic barriers (the road network, rivers etc.) or when hotspot centers are close to parks, lakes, mountains, etc. To address these limitations, a novel approach is proposed based on two ideas: cubic grid circle enumeration and a grid log likelihood ratio upper bound. A case study on real crime data shows that the proposed approach finds hotspots which cannot be discovered by the related work. Experimental results show that the proposed algorithm yields substantial computational savings compared to the related work.

Original languageEnglish (US)
Title of host publicationProceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015
EditorsXindong Wu, Alexander Tuzhilin, Hui Xiong, Jennifer G. Dy, Charu Aggarwal, Zhi-Hua Zhou, Peng Cui
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages10
ISBN (Electronic)9781467384926
StatePublished - Jan 29 2016
Event15th IEEE International Conference on Data Mining Workshop, ICDMW 2015 - Atlantic City, United States
Duration: Nov 14 2015Nov 17 2015

Publication series

NameProceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015


Other15th IEEE International Conference on Data Mining Workshop, ICDMW 2015
Country/TerritoryUnited States
CityAtlantic City

Bibliographical note

Funding Information:
This material is based upon work supported by the National Science Foundation under Grant No. 1029711, IIS-1320580, 0940818 and IIS-1218168, USDOD under Grant No. HM1582-08-1-0017, HM0210-13-1-0005, and University of Minnesota via U-Spatial. We would like to thank Kim Koffolt and the members of the University of Minnesota Spatial Computing Research Group for their comments.


  • Hotspot Detection
  • Spatial Data Mining
  • Spatial Scan Statistics


Dive into the research topics of 'Geographically Robust Hotspot Detection: A Summary of Results'. Together they form a unique fingerprint.

Cite this