The University of Arizona is running a research program to predict which students are most likely to drop out of college. The program tracks their student ID cards which are used by students at the library, rec center, student union, residence halls, and other facilities such as the convenience store, mail room, salon, and movie theater. Every time their ID cards are swiped at these locations a digital trace is left behind which is then tracked to predict the likelihood of students’ returning to campus after their freshman year.
The researchers are aiming to use this data to reduce drop out rates. The term drop out here refers to students who either transfer to another college or leave higher-education altogether.
The student ID cards have an embedded sensor and they are used at more than 700 locations on campus, including on vending machines. Sudha Ram, a professor of management information systems, is leading the research. So far, data has been gathered and analyzed over a three-year period, which was then used to create large networks in order to quantify patterns using machine learning algorithms.
“Of all the students who drop out at the end of the first year, with our social integration measures, we’re able to do a prediction at the end of the first 12 weeks of the semester with 85 to 90 percent recall,” Ram said, adding that this means out of the 2,000 students who drop out, they’re able to identify 1,800 of them.
Freshman retention is a major challenge for public universities across the country and the aim of this research is to reduce drop out rates by pinpoint to advisers which students need more support in order to stay on.
Filed in . Read more about Machine Learning. Source: uanews.arizona.edu