Landslide Forecast in Taiwan Based on Machine Learning in the GIS field - A Capstone Submitted to the University Honors Program

Aug 5, 2026

Upload by Miles

Landslides threaten life, property, and infrastructure in mountainous regions and are commonly triggered by intense rainfall, earthquakes, and changing water levels. Machine-learning forecasting typically relies on large datasets to capture statistical relations between past landslides and multiple variables, which can limit performance where data are scarce. This project evaluates an approach using only a few variables while leveraging spatial distribution and temporal trends. The study applies the method to Taiwan, aiming to support pre-planning and early warning for risk mitigation.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages20

Reading Time50 min

AccessFree to download

Contents

1. Introduction and Background

2. Data and Method

3. Results

4. Conclusion

5. Discussion

6. Acknowledgements

7. Reference

Northern Illinois University Huskie Commons | Honors Capstones | Undergraduate Research & Artistry | | --- | --- | | Spring 5-13-2023
Landslide Forecast in Taiwan Based on Machine Learning in the GIS field
Yi Shen
Follow this and additional works at: [https://huskiecommons.lib.niu.edu/studentengagement](https://huskiecommons.lib.niu.edu/studentengagement)honorscapstones
Part of the Earth Sciences Commons, and the Environmental Sciences Commons | | Recommended Citation Shen, Yi, "Landslide Forecast in Taiwan Based on Machine Learning in the GIS field" (2023) . Honors Capstones. 1465. [https://huskiecommons.lib.niu.edu/studentengagement-honorscapstones/1465](https://huskiecommons.lib.niu.edu/studentengagement-honorscapstones/1465) This Student Project is brought to you for free and open access by the Undergraduate Research & Artistry at Huskie Commons. It has been accepted for inclusion in Honors Capstones by an authorized administrator of Huskie Commons. For more information, please contact [jschumacher@niu.edu](jschumacher@niu.edu). NORTHERN ILLINOIS UNIVERSITY Landslide Forecast in Taiwan Based on Machine Learning in the GIS field. A Capstone Submitted to the University Honors Program In Partial Fulfillment of the Requirements of the Baccalaureate Degree With Honors Department Of Department of Earth, Atmosphere and Environment By Yi Shen DeKalb, Illinois 05/14/2023 Contents Abstract:................................................................................3 1. Introduction and Background .............................................................4 2. Data and Method .......................................................................6 3. Results .............................................................................. 10 4. Conclusion ........................................................................... 14 5. Discussion ........................................................................... 14 6. Acknowledgements .................................................................... 16 7. Reference ............................................................................ 17 Abstract: 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. Keywords: Landslide, Landslide Forecast, Taiwan, Machine Learning, GIS (Geographic Information System) , Remote Sensing, Spatial analysis. 1. Introduction and Background 1.1 Landslide Landslides refer to the movement of a mass of rock, debris, or earth down a slope (Cruden, 1991) . They can be triggered by various factors, including intense rainfall, earthquakes, and changes in water level (Dai, Lee, & Ngai, 2002) . Landslides have the potential to cause loss of life, property damage, and disruption of infrastructure, as well as psychological impacts such as trauma and stress (Kjekstad & Highland, 2009) . Therefore, early recognition and investigation of potential landslides are critical in preventing loss of life and damage to property and infrastructure (Ma, Mei,& Piccialli, 2020) . 1.2 Study Area In this research, our study area is Taiwan, which is prone to landslides due to the geology and geomorphology of the ar

FAQs

What problem does the project address?

The project addresses landslide risk in Taiwan and the need for more effective forecasting methods to support pre-planning and early warning.

How does the project improve forecasting with limited variables and data?

It uses only a few variables but incorporates both their spatial distribution and temporal trends to improve forecast accuracy under data constraints.

Which machine learning approach is central to the study?

The study is based on machine learning techniques for landslide prediction, including models such as decision trees and related methods discussed in the methodology.

Aug 5, 2026

Upload by Miles