Date Thesis Awarded


Access Type

Honors Thesis -- Access Restricted On-Campus Only

Degree Name

Bachelors of Arts (BA)




Ariel BenYishay

Committee Members

Philip Roessler

Daniel Runfola


Over the past decade, the number of individuals internally displaced by conflict (IDPs) has reached unprecedented levels. Humanitarian actors and first-responders face persistent information gaps in meeting the needs of these populations. Specifically, they face challenges in understanding where and how IDPs move after they are displaced, which is necessary to locate them in conflict-affected situations and provide them with life-saving assistance. In this paper, I propose a framework, using established machine-learning methods, to forecast the migration routes of these displaced populations (Chapter 1). In a case study of displacement in Yemen, my models predict 80% of IDPs' migration routes (Chapters 2 and 3). These results withstand significant robustness checks (Chapter 4). My findings suggest that IDPs are rational actors whose migration after displacement can be forecasted using a set of quantifiable factors. Machine learning methods can thus play a key role in developing innovative solutions to forecast migration routes, target assistance, and meet the needs of these vulnerable populations. I develop a tool that leverages my models to assist humanitarian organisations with these objectives (Chapter 5).

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

On-Campus Access Only