Research on finding patterns in large data sets wins award
A Wheaton College junior recently won second place in a national competition of undergraduate research in computer science with his development of a data analysis tool that could have far-ranging applications from business to medicine.
The winning scholarship by Eric Drewniak '11 developed and tested an algorithim for detecting patterns in large sets of data, an application that has grown in importance as major online firms such as Google, Amazon and Netflix have employed such methods to deliver core services of their business.
The math and computer science major's project was selected as the second-place finisher in the undergraduate research competition at the Special Interest Group on Computer Science Education (SIGSCE) conference, which was held in Milwaukee, Wisc. SIGSCE is the largest conference of computer science educators in the nation.
The appeal of the project, says Drewniak, lies in its applicability to a variety of problems.
"The data can be from any source: the closing prices of a stock over the course of a year, average temperature for a region over a decade, a buoy measuring water levels or temperatures in the ocean or a patient's heart rate," explained Drewniak, who conducted the research under the mentorship of Assistant Professor of Computer Science Tom Armstrong.
At the conference, Drewniak presented his research in an initial round of poster sessions, which placed him among the top five finishers. In the formal oral presentation that followed, he was awarded second place overall.
The Ludlow, Mass. native worked with Professor Armstrong on the research during the summer with support from the Mars Student-Faculty Research Partnership program at Wheaton College. He hopes to continue that work this summer, improving the speed of the algorithim through the use of parallel computing and expanding it to handle multiple-variable data sets.
"The diversity of data that can be examined makes the project particularly valuable to many fields. It is a way computer science can help benefit other fields," he explained. "Humans have a hard time finding patterns in large data sets and cannot constantly monitor incoming data."