Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Landslide Susceptibility Mapping of Chamoli (Uttarakhand) Using Random Forest Machine Learning Method
Landslides are among the most devastating natural calamities, causing significant changes in landscape morphology, compromising infrastructure, and posing grave threats to human lives. Developing accurate landslide susceptibility maps (LSMs) that indicate the likelihood of landslide occurrence in specific areas is paramount for effective environmental management, urban planning, and mitigating economic losses. Despite the importance of LSMs, existing research in data mining for landslide susceptibility has predominantly focused on small-scale case studies, typically examining single types of landslides. This paper presents a pioneering data mining approach to generate LSMs for the Chamoli district of Uttarakhand, India, a region with a heterogeneous landscape prone to multiple types of landslides. To address the complexity and scale of the problem, the researchers employed the Random Forest algorithm, a powerful and adaptable machine learning technique, to create both susceptibility and classification maps. A total of 13 causative factors were considered for the purposes of training and validation. The model underwent validation using the AUC-ROC method, yielding an AUC-ROC value of 82.7%. This value signifies that the predictive accuracy of the model is in close agreement with the real distribution of landslides across the study area. This accuracy is crucial for reliable risk assessment and informed decision-making. The innovative application of the Random Forest algorithm for landslide susceptibility mapping in a large and heterogeneous region like Chamoli sets a promising precedent. The findings of the study have significant implications for enhancing landslide risk management strategies in similar regions globally, facilitating proactive measures to safeguard lives and assets in Landslide-prone areas.
Landslide Susceptibility Mapping of Chamoli (Uttarakhand) Using Random Forest Machine Learning Method
Landslides are among the most devastating natural calamities, causing significant changes in landscape morphology, compromising infrastructure, and posing grave threats to human lives. Developing accurate landslide susceptibility maps (LSMs) that indicate the likelihood of landslide occurrence in specific areas is paramount for effective environmental management, urban planning, and mitigating economic losses. Despite the importance of LSMs, existing research in data mining for landslide susceptibility has predominantly focused on small-scale case studies, typically examining single types of landslides. This paper presents a pioneering data mining approach to generate LSMs for the Chamoli district of Uttarakhand, India, a region with a heterogeneous landscape prone to multiple types of landslides. To address the complexity and scale of the problem, the researchers employed the Random Forest algorithm, a powerful and adaptable machine learning technique, to create both susceptibility and classification maps. A total of 13 causative factors were considered for the purposes of training and validation. The model underwent validation using the AUC-ROC method, yielding an AUC-ROC value of 82.7%. This value signifies that the predictive accuracy of the model is in close agreement with the real distribution of landslides across the study area. This accuracy is crucial for reliable risk assessment and informed decision-making. The innovative application of the Random Forest algorithm for landslide susceptibility mapping in a large and heterogeneous region like Chamoli sets a promising precedent. The findings of the study have significant implications for enhancing landslide risk management strategies in similar regions globally, facilitating proactive measures to safeguard lives and assets in Landslide-prone areas.
Landslide Susceptibility Mapping of Chamoli (Uttarakhand) Using Random Forest Machine Learning Method
Lecture Notes in Civil Engineering
Hazarika, Hemanta (Herausgeber:in) / Haigh, Stuart Kenneth (Herausgeber:in) / Chaudhary, Babloo (Herausgeber:in) / Murai, Masanori (Herausgeber:in) / Manandhar, Suman (Herausgeber:in) / Mittal, Amogh (Autor:in) / Gupta, Kunal (Autor:in) / Satyam, Neelima (Autor:in)
International Conference on Construction Resources for Environmentally Sustainable Technologies ; 2023 ; Fukuoka, Japan
04.05.2024
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch