Tackling the Opioid Crisis with Machine Learning Models

Project Title

Tackling the Opioid Crisis: Machine Learning Models for Predicting and Identifying High-Risk Counties

Faculty Mentor(s)

Project Description

Our research is focused on developing the best-performing machine learning model at predicting the number of opioid overdose deaths across U.S counties and which counties will be the top decile of opioid overdose death rates for us to be aware of counties at high risk.

The significance of our research lies in its application in public health. Machine learning models can assist in predicting death counts and identifying counties at high risk of opioid overdose deaths in the future so that public policy makers are informed how to take measures to combat the overall impact of the opioid crisis.

I got more involved with research through the STRONG program. Doing research is an opportunity to learn what you are passionate about in a less structured way.

How long have you been conducting research?

I began conducting research in collaboration with Professor Adam Eck during the Winter Term of 2023. Our primary focus is on applying Machine Learning to address critical challenges in global health.

Please include a brief description of your research project

In the U.S, previous studies have shown that the number of drug overdose deaths has been increasing at an alarming rate. Approximately two thirds of the deaths are associated with opioids. Urgent action is required to save lives and also reduce financial burden on the health system. Our research is focused on developing the best-performing machine learning model at predicting the number of opioid overdose deaths across U.S counties and which counties will be the top decile of opioid overdose death rates for us to be aware of counties at high risk. Using relevant datasets, we train, validate, and test various Machine Learning models such as linear regression, gradient boosting, support vector machine, random forest, neural network, and decision trees. We evaluate performance of the machine Learning models using a set of justified performance metrics to determine the optimal model.

Please include a brief summary (the elevator speech) of your research project

Application of Machine learning in predicting future deaths due to opioid overdose. By leveraging these predictions, we can effectively identify counties and states that are at high risk of opioid overdose.

Why is your research important?

The significance of our research lies in its application in public health. Machine learning models can assist in predicting death counts and identifying counties at high risk of opioid overdose deaths in the future so that public policy makers are informed how to take measures to combat the overall impact of the opioid crisis.

What does the process of doing your research look like?

In the course of my research, I start the week by setting clear goals and objectives, which I review at the end of the week to assess progress. Regular weekly meetings with my mentor play a crucial role, where I present my findings and engage in discussions about the broader aspects of the research.

What knowledge has your research contributed to your field?

In all our experiments, our results show that random forest models outperformed all other models across all performance measures. The models’ performance also got better with more inclusion of historical data even though the difference was not statistically significant.

In what ways have you showcased your research thus far?

I have presented in the Winter Term presentations events, Oberlin Undergraduate Research Symposium, and Oberlin Summer Research Institute Symposium.

How did you get involved in research? What drove you to seek out research experiences in college?

I got more involved with research through the STRONG program. Doing research is an opportunity to learn what you are passionate about in a less structured way.

What is your favorite aspect of the research process?

The most enjoyable aspect was discovering new directions and unexpected results, which can be quite surprising at times. Additionally, engaging in meaningful discussions with my mentor that enriched the overall research experience.

How has working with your mentor impacted the development of your research project? How has it impacted you as a researcher?

Through my mentor’s guidance, I gained insights into conducting an extensive literature review, leading to better formulation of research questions. Regularly presenting to my mentor has greatly improved my public speaking skills, making me a confident and effective researcher.

How has the research you’ve conducted contributed to your professional or academic development?

Conducting research honed my time-management and data-analysis. Collaborating with my mentor has provided valuable experience in teamwork. Moreover, this exposure to various skills has guided me in making informed decisions regarding the elective courses I can pursue within my Computer Science Major.

What advice would you give to a younger student wanting to get involved in research in your field?

Don’t hesitate to reach out to the professors and other students conducting research to know more about the research they are doing. I would also suggest taking advantage of opportunities like winter term and summer research to get started with research.