A Comparison of Markov Random Field and Spatial Regression Models for Mining Geospatial Data

Sanjay Chawla, Shashi Shekhar, Weili Wu

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

1 Scopus citations

Abstract

A spatial metric, which can be used to systematically calculate the regularization parameter in an MRF formulation of a spatial classification problem was proposed. The standard way to measure classification accuracy is to calculate the percentage of correctly classified objects. Spatial accuracy achieved by classical regression, Spatial Autoregressive Regression (SAR), and Markov Random Field (MRF_hMETIS) was compared using Average Distance to Nearest Prediction (ANDP). It was observed that spatial regression takes two orders of magnitude more computation time relative to MRF_hMETIS approach using the public domain code.

Original languageEnglish (US)
Title of host publicationProceedings of the 6th Joint Conference on Information Sciences, JCIS 2002
EditorsJ.H. Caulfield, S.H. Chen, H.D. Cheng, R. Duro, J.H. Caufield, S.H. Chen, H.D. Cheng, R. Duro, V. Honavar
Pages245-250
Number of pages6
StatePublished - Dec 1 2002
EventProceedings of the 6th Joint Conference on Information Sciences, JCIS 2002 - Research Triange Park, NC, United States
Duration: Mar 8 2002Mar 13 2002

Publication series

NameProceedings of the Joint Conference on Information Sciences
Volume6

Other

OtherProceedings of the 6th Joint Conference on Information Sciences, JCIS 2002
CountryUnited States
CityResearch Triange Park, NC
Period3/8/023/13/02

Fingerprint Dive into the research topics of 'A Comparison of Markov Random Field and Spatial Regression Models for Mining Geospatial Data'. Together they form a unique fingerprint.

Cite this