Forecasting student success with machine learning
Drew Wham, a data scientist for Penn State’s Education Technology Services, will deliver a talk entitled “Using Machine Learning to Forecast Student Outcomes.” The talk will occur Oct. 5 from 1:30 to 3 p.m. in 233A HUB-Robeson Center.
This event is free and open to the public. The lecture is sponsored by the Institute for CyberScience (ICS) as part of the ICS CyberScience Seminars, a series of talks on cutting-edge topics of interest to the cyberscience research community at Penn State.
Wham is working on a computational method for modeling and predicting college students’ end-of-semester GPAs, as well as their probability of withdrawal, before the beginning of the semester. Some students struggle with the courses they have selected for a variety of reasons: for example, they may be under-prepared for an individual course, or the combination of courses may present too large a course load. Wham’s method can predict these struggles early.
The idea is that, if academic advisers and instructors could identify students who are strugging early on, they could give guidance in course scheduling and sequencing. They would also be able to provide resources that help students improve their performance, leading to better grades, higher retention rates and happier students.
In his talk, Wham will discuss how his model works and ways it can be applied to improve student outcomes.
Space is limited, so please reserve a seat at the seminar by October 2. The event includes Wham’s talk, an extended question-and-answer session, and time to socialize. Refreshments will be served.
ICS CyberScience Seminars explore a wide range of interdisciplinary topics. Check out the full slate of speakers for 2017-18.
Prior to joining Education Technology Services, Wham earned his PhD in biology at Penn State. He specialized in using Bayesian modeling, machine learning and statistical programming to solve problems in genetics and bioinformatics. Some of his major projects have focused on detecting and naming new species, predicting the number of unique individuals in clonal populations and inferring the rate of clonal reproduction. His research focuses on problems that involve hidden groups and data that comes from individuals that are related by observable or un-observable networks.