B.S. (Bachelor of Science)
Department of Geographic and Atmospheric Sciences
Landslides can pose a significant risk to life, property, and infrastructure in mountainous regions, and can be triggered by various factors, including intense rainfall, earthquakes, and water level changes. Machine learning is commonly used to forecast landslides, based on statistical relationships between past landslides and multiple variables to create a general forecasting model. However, these models often require large amounts of data to achieve accurate results. This project aims to use only a few variables but take advantage of both their spatial distribution and temporal trends to improve the accuracy of landslide forecasts. This approach is tested in Taiwan, a region prone to landslides, and could be beneficial for pre-planning and early warning systems in areas with limited data to mitigate landslide risks.
Shen, Yi, "Landslide Forecast in Taiwan Based on Machine Learning in the GIS field" (2023). Honors Capstones. 1465.