Utilization of Classical and Quantum Machine Learning-Based Models to Study the Human Migration Dynamics

Author

Thilanka Munasinghe, University at Albany, State University of New YorkFollow

ORCID

https://orcid.org/0000-0002-0911-750X

Date of Award

Spring 2025

Language

English

Embargo Period

5-1-2027

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School/Department

Department of Information Science and Technology

Program

Information Science

First Advisor

Kimberly A. Cornell

Second Advisor

George Berg

Committee Members

James A. Hendler, Jennifer C. Wei

Keywords

Quantum Machine Learning, Classical Machine Learning, Human Migration, Socioeconomics, Environmental, Demographic

Subject Categories

Data Science | Multivariate Analysis | Numerical Analysis and Scientific Computing | Other Computer Sciences | Science and Technology Studies | Statistical Models | Theory and Algorithms | Urban Studies and Planning

Abstract

Socioeconomic, demographic, and climate-related extreme weather events have significantly influenced human dynamics such as short-term mobility and migration across the world. This dissertation investigates the interplay between environmental and socioeconomic-demographic factors and human migration using a data-driven approach. The study focused on implementing machine learning-based predictive models for county-to-county residential migration in the state of Texas, in the United States of America, leveraging both classical and quantum machine learning techniques.

The research is structured into two primary objectives: (1) to integrate and analyze environmental, socioeconomic, and migration datasets to understand what could influence human migration, and (2) to evaluate the feasibility of quantum machine learning models for data analytics compared to traditional classical machine learning approaches. Datasets include county-to-county residential mobility data from the US Census Bureau’s socioeconomic and demographic indicator data, and Earth observational environmental and weather data from NASA.

Classical machine learning models, including Support Vector Machines (SVM), Logistic Regression, Extreme Gradient Boosting (XGBoost), and Huber regression method, were employed to conduct data analytics for county-to-county migration. This study explored the feasibility of quantum machine learning-based classification models, such as the Quantum Support Vector Classifier (QSVC) and the Variational Quantum Classifier (VQC), implemented within the IBM Qiskit quantum computing framework.

However, current quantum hardware limitations pose practical challenges when using a large number of features to train the model compared to classical machine learning approaches. Experimental results indicate that classical machine learn- ing models offer strong predictive power compared to their quantum counterparts, which utilize current quantum computing technology. Quantum machine learning models are still in the early stages of development as of writing this dissertation.

Our study revealed the practical limitations of using a larger number of features developed using real-world data on a quantum computing environment to study real-world phenomena of human migration. Quantum computers have a path to exhibit potential advantages as technology develops. The findings from our exploratory analysis and machine learning model development exercises contribute to the understanding of county-to-county residential migration, which is shaped by socioeconomic indicators and environmental factors. Furthermore, this study demonstrates the applicability and limitations of both classical and quantum machine learning in data analytics for environmental and social systems, providing insights on how to develop quantum machine learning applications in similar domains.

License

This work is licensed under the University at Albany Standard Author Agreement.

Recommended Citation

Munasinghe, Thilanka, "Utilization of Classical and Quantum Machine Learning-Based Models to Study the Human Migration Dynamics" (2025). Electronic Theses & Dissertations (2024 - present). 186.
https://scholarsarchive.library.albany.edu/etd/186