Institute Team Extracts Meaningful Patterns

Main page content

A team from the Institute of Technology has reached the final round in an international data-mining competition.

UW Tacoma Institute of Technology Professor Martine De Cock(11/14/16 Update: the UW Tacoma team won 2nd place in the 2016 AMIA KDDM Data Mining Competition.)

A group of UW Tacoma Institute of Technology faculty and students are on a team that has reached the final round in an international data mining competition.

The team includes:
Prathyusha Mandagani, graduate student, MSCSS program
Shaun Coleman, undergraduate student, CSS program
Anam Zahid, graduate student, MSCSS program
Anne Pugel Ehlers, general surgery research resident program, UW School of Medicine
Senjuti Basu Roy, assistant professor, New Jersey Institute of Technology (formerly with UW Tacoma)
Martine De Cock, professor, UW Tacoma Institute of Technology

The team was faced with the challenge of predicting which hospital surgery patients would be more likely to develop post-operative hospital-acquired infections. Surgical site infections are estimated to represent 17% of all hospital-acquired infections, prolonging hospital stays and increasing healthcare costs.

Help Make a Difference

You can help empower students who dream of changing the world with a degree from UW Tacoma’s Institute of Technology.

Make your gift today as part of the Campaign for UW Tacoma, and help build a greater Tacoma and a greater world.

Make a Gift

Competition teams were provided a data set of about 7,725 patients from a hospital in Norway, along with a total of more than 4.5 million of their blood tests. The team of students and faculty from the Institute of Technology’s Center for Data Science, working together with Anne P. Ehlers from UW Medicine, had to develop, as per the competition instructions, “[a] process of extracting meaningful patterns from [the data], using automated computational and statistical tools and techniques on large datasets.”

The extraction of meaningful patterns happens through two processes called knowledge discovery and data mining. UW Tacoma is developing an international reputation for its work developing these techniques, especially as they apply to the field of healthcare.

Team leader Martine De Cock said, “Within a week after the competition was launched, we had our first—quite poor!—results. Then, week after week, the students refined our techniques, and the results steadily improved. All the students on the team volunteered their time, fitting the project into their already very busy schedules of coursework, internships and taking care of their families.”

Other work at UW Tacoma’s Center for Data Science using these same techniques has resulted in the first corporate startup to be launched based on technology developed at UW Tacoma, a company called KenSci, which has developed what it calls “the world’s first machine learning app store for healthcare.”

Three teams, selected from a large field, reached the final round in the data mining competition, managed by the American Medical Informatics Association. The three teams, including UW Tacoma’s, will convene in Chicago in November to present their solutions to the challenge, at which time 1st, 2nd and 3rd place winners will be announced. All three teams’ papers describing their solutions will be published in the International Journal of Big Data and Analytics in Healthcare.

Section: 
Written by: 
John Burkhardt / October 19, 2016
Media contact: 

John Burkhardt, UW Tacoma Communications, 253-692-4536 or johnbjr@uw.edu