Assessment of Landslide Susceptibility Using Machine Learning Techniques in a Mountainous Region

Thesis Abstract

Abstract
Landslides pose a significant threat to communities residing in mountainous regions, leading to loss of lives, property damage, and disruption of infrastructure. The assessment of landslide susceptibility is crucial for effective risk management and mitigation strategies. This thesis investigates the application of machine learning techniques in assessing landslide susceptibility in a mountainous region. The study focuses on leveraging the power of machine learning algorithms to analyze various factors contributing to landslide occurrences and predict susceptible areas with high accuracy. The research methodology involves the collection of geospatial data, including topographic attributes, land cover types, soil properties, and historical landslide events. Machine learning models such as Random Forest, Support Vector Machine, and Logistic Regression are applied to develop landslide susceptibility maps based on these input variables. The performance of the models is evaluated using statistical measures such as sensitivity, specificity, and area under the curve. The literature review provides a comprehensive overview of previous studies on landslide susceptibility assessment, machine learning applications in geosciences, and the significance of incorporating various environmental factors in landslide modeling. The review highlights the advantages and limitations of machine learning techniques compared to traditional methods in landslide susceptibility analysis. The findings of the study reveal the effectiveness of machine learning algorithms in predicting landslide susceptibility in the mountainous region. The developed susceptibility maps exhibit high accuracy in identifying areas prone to landslides based on the input variables considered. The results also demonstrate the importance of integrating multiple environmental factors to enhance the predictive capability of the models. The discussion chapter delves into the implications of the research findings and discusses the practical applications of the developed landslide susceptibility maps in land use planning, disaster risk reduction, and emergency response strategies. The limitations of the study, such as data availability and model complexity, are also addressed, providing insights for future research directions. In conclusion, the assessment of landslide susceptibility using machine learning techniques offers a promising approach for enhancing landslide risk assessment and management in mountainous regions. The study contributes to the growing body of knowledge on geospatial analysis and highlights the potential of machine learning in addressing complex geohazard challenges. Recommendations for further research include the refinement of models, the incorporation of real-time data, and the validation of results through field investigations. Keywords Landslide susceptibility, Machine learning, Geospatial analysis, Risk assessment, Mountainous region, Geohazards.