Spatial classification and prediction models for geospatial data mining

Shashi Shekhar, Ranga Raju Vatsavai, Sanjay Chawla

Research output: Chapter in Book/Report/Conference proceedingChapter

3 Citations (Scopus)

Abstract

Widespread use of spatial databases [42], an important subclass of multimedia databases, is leading to an increased interest in mining interesting and useful but implicit spatial patterns [23, 29, 18, 40]. Traditional data mining algorithms [1] often make assumptions (e.g., independent, identical distributions) which violate Tobler’s rst law of geography: everything is related to everything else but nearby things are more related than distant things [45]. In other words, the values of attributes of nearby spatial objects tend to systematically affect each other. In spatial statistics, an area within statistics devoted to the analysis of spatial data, this is called spatial autocorrelation [12]. Knowledge discovery techniques that ignore spatial autocorrelation typically perform poorly in the presence of spatial data. The simplest way to model spatial dependence is through spatial covariance. Often the spatial dependencies arise due to the inherent characteristics of the phenomena under study, but in particular they arise due to the fact that imaging sensors have better resolution than object size. For example, remote sensing satellites have resolutions ranging from 30 m (e.g., Enhanced Thematic Mapper of Landsat 7 satellite of NASA) to 1 m (e.g., IKONOS satellite from SpaceImaging), while the objects under study (e.g., urban, forest, water) are much bigger than 30 m. As a result, the per-pixel-based classi- ers, which do not take spatial context into account, often produce classied images with salt and pepper noise. These classiers also suffer in terms of classication accuracy.

Original languageEnglish
Title of host publicationGeographic Data Mining and Knowledge Discovery, Second Edition
PublisherCRC Press
Pages117-148
Number of pages32
ISBN (Electronic)9781420073980
ISBN (Print)9781420073973
DOIs
Publication statusPublished - 1 Jan 2009

Fingerprint

data mining
Data mining
Satellites
Autocorrelation
spatial data
autocorrelation
prediction
Statistics
IKONOS
multimedia
Landsat
NASA
Remote sensing
pixel
Pixels
Salts
sensor
salt
remote sensing
Imaging techniques

ASJC Scopus subject areas

  • Computer Science(all)
  • Engineering(all)
  • Earth and Planetary Sciences(all)

Cite this

Shekhar, S., Vatsavai, R. R., & Chawla, S. (2009). Spatial classification and prediction models for geospatial data mining. In Geographic Data Mining and Knowledge Discovery, Second Edition (pp. 117-148). CRC Press. https://doi.org/10.1201/9781420073980

Spatial classification and prediction models for geospatial data mining. / Shekhar, Shashi; Vatsavai, Ranga Raju; Chawla, Sanjay.

Geographic Data Mining and Knowledge Discovery, Second Edition. CRC Press, 2009. p. 117-148.

Research output: Chapter in Book/Report/Conference proceedingChapter

Shekhar, S, Vatsavai, RR & Chawla, S 2009, Spatial classification and prediction models for geospatial data mining. in Geographic Data Mining and Knowledge Discovery, Second Edition. CRC Press, pp. 117-148. https://doi.org/10.1201/9781420073980
Shekhar S, Vatsavai RR, Chawla S. Spatial classification and prediction models for geospatial data mining. In Geographic Data Mining and Knowledge Discovery, Second Edition. CRC Press. 2009. p. 117-148 https://doi.org/10.1201/9781420073980
Shekhar, Shashi ; Vatsavai, Ranga Raju ; Chawla, Sanjay. / Spatial classification and prediction models for geospatial data mining. Geographic Data Mining and Knowledge Discovery, Second Edition. CRC Press, 2009. pp. 117-148
@inbook{0d80e523ad6341c2833db4547b2eaebb,
title = "Spatial classification and prediction models for geospatial data mining",
abstract = "Widespread use of spatial databases [42], an important subclass of multimedia databases, is leading to an increased interest in mining interesting and useful but implicit spatial patterns [23, 29, 18, 40]. Traditional data mining algorithms [1] often make assumptions (e.g., independent, identical distributions) which violate Tobler’s rst law of geography: everything is related to everything else but nearby things are more related than distant things [45]. In other words, the values of attributes of nearby spatial objects tend to systematically affect each other. In spatial statistics, an area within statistics devoted to the analysis of spatial data, this is called spatial autocorrelation [12]. Knowledge discovery techniques that ignore spatial autocorrelation typically perform poorly in the presence of spatial data. The simplest way to model spatial dependence is through spatial covariance. Often the spatial dependencies arise due to the inherent characteristics of the phenomena under study, but in particular they arise due to the fact that imaging sensors have better resolution than object size. For example, remote sensing satellites have resolutions ranging from 30 m (e.g., Enhanced Thematic Mapper of Landsat 7 satellite of NASA) to 1 m (e.g., IKONOS satellite from SpaceImaging), while the objects under study (e.g., urban, forest, water) are much bigger than 30 m. As a result, the per-pixel-based classi- ers, which do not take spatial context into account, often produce classied images with salt and pepper noise. These classiers also suffer in terms of classication accuracy.",
author = "Shashi Shekhar and Vatsavai, {Ranga Raju} and Sanjay Chawla",
year = "2009",
month = "1",
day = "1",
doi = "10.1201/9781420073980",
language = "English",
isbn = "9781420073973",
pages = "117--148",
booktitle = "Geographic Data Mining and Knowledge Discovery, Second Edition",
publisher = "CRC Press",

}

TY - CHAP

T1 - Spatial classification and prediction models for geospatial data mining

AU - Shekhar, Shashi

AU - Vatsavai, Ranga Raju

AU - Chawla, Sanjay

PY - 2009/1/1

Y1 - 2009/1/1

N2 - Widespread use of spatial databases [42], an important subclass of multimedia databases, is leading to an increased interest in mining interesting and useful but implicit spatial patterns [23, 29, 18, 40]. Traditional data mining algorithms [1] often make assumptions (e.g., independent, identical distributions) which violate Tobler’s rst law of geography: everything is related to everything else but nearby things are more related than distant things [45]. In other words, the values of attributes of nearby spatial objects tend to systematically affect each other. In spatial statistics, an area within statistics devoted to the analysis of spatial data, this is called spatial autocorrelation [12]. Knowledge discovery techniques that ignore spatial autocorrelation typically perform poorly in the presence of spatial data. The simplest way to model spatial dependence is through spatial covariance. Often the spatial dependencies arise due to the inherent characteristics of the phenomena under study, but in particular they arise due to the fact that imaging sensors have better resolution than object size. For example, remote sensing satellites have resolutions ranging from 30 m (e.g., Enhanced Thematic Mapper of Landsat 7 satellite of NASA) to 1 m (e.g., IKONOS satellite from SpaceImaging), while the objects under study (e.g., urban, forest, water) are much bigger than 30 m. As a result, the per-pixel-based classi- ers, which do not take spatial context into account, often produce classied images with salt and pepper noise. These classiers also suffer in terms of classication accuracy.

AB - Widespread use of spatial databases [42], an important subclass of multimedia databases, is leading to an increased interest in mining interesting and useful but implicit spatial patterns [23, 29, 18, 40]. Traditional data mining algorithms [1] often make assumptions (e.g., independent, identical distributions) which violate Tobler’s rst law of geography: everything is related to everything else but nearby things are more related than distant things [45]. In other words, the values of attributes of nearby spatial objects tend to systematically affect each other. In spatial statistics, an area within statistics devoted to the analysis of spatial data, this is called spatial autocorrelation [12]. Knowledge discovery techniques that ignore spatial autocorrelation typically perform poorly in the presence of spatial data. The simplest way to model spatial dependence is through spatial covariance. Often the spatial dependencies arise due to the inherent characteristics of the phenomena under study, but in particular they arise due to the fact that imaging sensors have better resolution than object size. For example, remote sensing satellites have resolutions ranging from 30 m (e.g., Enhanced Thematic Mapper of Landsat 7 satellite of NASA) to 1 m (e.g., IKONOS satellite from SpaceImaging), while the objects under study (e.g., urban, forest, water) are much bigger than 30 m. As a result, the per-pixel-based classi- ers, which do not take spatial context into account, often produce classied images with salt and pepper noise. These classiers also suffer in terms of classication accuracy.

UR - http://www.scopus.com/inward/record.url?scp=74049153100&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=74049153100&partnerID=8YFLogxK

U2 - 10.1201/9781420073980

DO - 10.1201/9781420073980

M3 - Chapter

SN - 9781420073973

SP - 117

EP - 148

BT - Geographic Data Mining and Knowledge Discovery, Second Edition

PB - CRC Press

ER -