Undergrad Statistics Major Explores Large Data Sets Through Machine Learning

October 21, 2016

For Michael Todd Young, a senior statistics major, the opportunity to apply classroom knowledge to research presented itself in the form of an analytics project predicting stock prices.

“During my time in the undergraduate statistics program I’ve become fascinated by machine learning,” Young said. “It blends my favorite subjects of computer science and statistics in a playful and profound way.”

Machine learning allows students to readily explore the patterns within data. “With a little luck,” Young said, “We can learn something that would not have been possible otherwise.”

Young hopes to submit an academic paper for publication with his adviser, Michel Ballings, assistant professor of business analytics. It will present data-driven models for predicting the direction in which stock prices will move.

According to Ballings, Young is learning LaTex, a document preparation system, improving his Linux and R programming skills, learning to position an academic paper and persevering despite the challenges of an expansive machine learning research project.

Ballings said he wants to provide Young with an opportunity to engage in scientific practice beyond the classroom setting.

“This work ties together a lot of the material he has learned in his studies,” Ballings said. “Machine learning is an iterative process of trial and error that is very unstructured. He has had to acquire these new skills.”

Young said Ballings’ notion that machine learning represents a novel approach to the scientific method resonates with him.

“In essence, we are constantly racing through the scientific method,” Young said. “It feels like we’re constantly working on the edge of the unknown, and for a curious person like me that’s very exciting. Machine learning allows us create working models in places where we have no underlying theory.”