Comparison of major statistical methods and their combination using matrix validation for landslide susceptibility mapping

A. Q. Akbar, Guangqi Chen

Research output: Contribution to journalArticle

Abstract

Landslide risk exists with the mountain regions and every year creates a great life and financial losses. To prevent the disaster, numbers of statistical methods have been proposed, but it is still unclear which one is more accurate and yet very few studies proposes a reliable method. Therefore, this study aims to compare the commonly used bivariate statistical method and multivariate statistical methods and their combination to achieve higher accuracy for landslide susceptibility map. Moreover, the classification used for landslide susceptibility mapping is associated with errors, which affects the accuracy of the analysis. In this study, new tool was designed to reduce the classification. To implement this study, a landslide susceptibility maps were created Kabul city. The result proposes that the new designed tool is a good way not only to reduce the classification error by defining the critical thresholds for the classifications. Moreover, all of the statistical methodologies is giving and acceptable result but the combination bivariate and multivariate statistical methods increase the accuracy of the analysis and they are complimentary to each other.

Original languageEnglish
Pages (from-to)401-412
Number of pages12
Journallowland technology international
Volume20
Issue number3
Publication statusPublished - Jan 1 2018

Fingerprint

Landslides
Statistical methods
Disasters

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Comparison of major statistical methods and their combination using matrix validation for landslide susceptibility mapping. / Akbar, A. Q.; Chen, Guangqi.

In: lowland technology international, Vol. 20, No. 3, 01.01.2018, p. 401-412.

Research output: Contribution to journalArticle

@article{dd55be759a4145f38157314dbf0224a0,
title = "Comparison of major statistical methods and their combination using matrix validation for landslide susceptibility mapping",
abstract = "Landslide risk exists with the mountain regions and every year creates a great life and financial losses. To prevent the disaster, numbers of statistical methods have been proposed, but it is still unclear which one is more accurate and yet very few studies proposes a reliable method. Therefore, this study aims to compare the commonly used bivariate statistical method and multivariate statistical methods and their combination to achieve higher accuracy for landslide susceptibility map. Moreover, the classification used for landslide susceptibility mapping is associated with errors, which affects the accuracy of the analysis. In this study, new tool was designed to reduce the classification. To implement this study, a landslide susceptibility maps were created Kabul city. The result proposes that the new designed tool is a good way not only to reduce the classification error by defining the critical thresholds for the classifications. Moreover, all of the statistical methodologies is giving and acceptable result but the combination bivariate and multivariate statistical methods increase the accuracy of the analysis and they are complimentary to each other.",
author = "Akbar, {A. Q.} and Guangqi Chen",
year = "2018",
month = "1",
day = "1",
language = "English",
volume = "20",
pages = "401--412",
journal = "Lowland Technology International",
issn = "1344-9656",
publisher = "International Association of Lowland Technology",
number = "3",

}

TY - JOUR

T1 - Comparison of major statistical methods and their combination using matrix validation for landslide susceptibility mapping

AU - Akbar, A. Q.

AU - Chen, Guangqi

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Landslide risk exists with the mountain regions and every year creates a great life and financial losses. To prevent the disaster, numbers of statistical methods have been proposed, but it is still unclear which one is more accurate and yet very few studies proposes a reliable method. Therefore, this study aims to compare the commonly used bivariate statistical method and multivariate statistical methods and their combination to achieve higher accuracy for landslide susceptibility map. Moreover, the classification used for landslide susceptibility mapping is associated with errors, which affects the accuracy of the analysis. In this study, new tool was designed to reduce the classification. To implement this study, a landslide susceptibility maps were created Kabul city. The result proposes that the new designed tool is a good way not only to reduce the classification error by defining the critical thresholds for the classifications. Moreover, all of the statistical methodologies is giving and acceptable result but the combination bivariate and multivariate statistical methods increase the accuracy of the analysis and they are complimentary to each other.

AB - Landslide risk exists with the mountain regions and every year creates a great life and financial losses. To prevent the disaster, numbers of statistical methods have been proposed, but it is still unclear which one is more accurate and yet very few studies proposes a reliable method. Therefore, this study aims to compare the commonly used bivariate statistical method and multivariate statistical methods and their combination to achieve higher accuracy for landslide susceptibility map. Moreover, the classification used for landslide susceptibility mapping is associated with errors, which affects the accuracy of the analysis. In this study, new tool was designed to reduce the classification. To implement this study, a landslide susceptibility maps were created Kabul city. The result proposes that the new designed tool is a good way not only to reduce the classification error by defining the critical thresholds for the classifications. Moreover, all of the statistical methodologies is giving and acceptable result but the combination bivariate and multivariate statistical methods increase the accuracy of the analysis and they are complimentary to each other.

UR - http://www.scopus.com/inward/record.url?scp=85066256553&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066256553&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:85066256553

VL - 20

SP - 401

EP - 412

JO - Lowland Technology International

JF - Lowland Technology International

SN - 1344-9656

IS - 3

ER -