Applying predictive analytics in ESP courses based on students' writing
Applying predictive analytics in ESP courses based on students' writing

Summary/Abstract: Fierce competition of student admission to strong reputation higher education institutions as well as getting employment in respected companies has built the need to profile potential candidates and select those that best meet one's requirements. Our research aimed to predict students' success potential, i.e. grades achieved in their English for Specific Purposes courses. Total of 292 students studying at the Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Croatia and 150 students studying at the Faculty of Material Science and Technology in Trnava, Slovakia voluntarily participated in the research by submitting their short biographies. The biographies were analyzed with the software for computational analysis (LIWC) whose output, in the form of raw numbers, was entered in an application specifically designed for this purpose. Based on the input data, the application calculated students' grades in the aforementioned courses which were later compared to the students' actual grades. The research has proven the application's high efficiency since it correctly predicted students' grades in 75% of the cases and in additional 12%, the grades were approximately predicted, i.e. the predicted grade was one grade higher/lower than the actual grade.

• Details
• Contents