AI in Teaching: Predictive Analytics
Predictive analytics, as the name implies, aims to predict the future based on historical data in order to eliminate or be prepared to undesired events. Some of the most popular applications are including but not limited to demand forecast and customer retention for retailers, stock market prediction for investors and various financial predictions of banks. What if an algorithm can predict your grade from a class based on your demographics and background information?
As crazy as it sounds, I don’t think that’s not an impossible scenario – especially from the eyes of education data mining researchers. Student retention is an important goal for universities because each student dropping out of the programme has a big opportunity cost for the institution – unused resources. U.S. Higher Ed institutions are recently having all-time highest student drop out rates. Recent trends show that only half of all students leave college with a diploma. Some of the schools have 47% student retention just in freshman year (source).
There is something this data tells us: we should change things and we should improve our education system. In the age of big data, it is very important for universities to be data-informed. Perceivant is a tech company aiming to serve exactly this purpose: struggle detection. As a student starts having an unusual decrease in class performance, the algorithm spots an increase in likelihood of failure and gives an early warning to the educator. This way educator finds the opportunity to accommodate and help the student.
Struggle detection is not the only cool feature of Perceivant’s learning management system. Another good feature is the adaptive learning practices which helps educator to set the teaching style based on the current student engagement. Using this platform, Kennesaw State University decreased dropout rates by 50 percent along with increased GPAs of the students (source).
As the data and AI technologies become more available for innovators from different disciplines, I believe that the field of Educational Data Mining will focus on this topic even more. Who knows – with the overnight online education shift, we might make great use of predictive analytics as we might need to be more reliant on the data.
April 4, 2020 @ 1:20 am
AI is a new world that could be applied in many fields. Education definitely is one of these fields and I like your idea in the blog that data tell us something. When studying econometric as a master student, I was always suggested to let the data talk. Data-informed and data-driven will surely play an important role in future education. Maybe it would be better if you can provide a more specific example of how AI changes higher education.
April 4, 2020 @ 7:00 pm
I think this is really interesting. One of my friends tells a story of a professor that would assign final grades after the first 60% of the class assignments had been graded because in all their years of teaching that class no student had ever been able to earn something different after the 60% point. I think that is crazy but I imagine an AI system would do exactly that. I’m not sure how I feel about that.