Technology to forecast educational results (on the example of “Chemistry” studied by students of construction and engineering university)

The purpose of this article is to develop a forecasting technology that makes it possible to predict the learning outcomes of civil engineers’ students studying “Chemistry” during the first weeks of study. The method of cluster analysis of k-means was chosen as a diagnostic tool. The participants were the students of 19 academic groups of full-time bachelor's degree program. For the experiment, the results of current control group (with an accuracy of 0.5 points) were collected from 298 first-year students that were asked to predict (on a 100-point scale) their score for the course being studied. The developed technology allows predicting learning grades with satisfactory accuracy equal to 86.24%. Eight hypotheses were tested; the results can be divided into three clusters of students: those who received a positive score; those who achieved a satisfactory result; students who are required to pass the exam to obtain a “satisfactory” grade. The study contributes to the development of computer pedagogy. It can be useful for teachers (for recommending to visit consultations to students who cannot achieve good results on their own, for a more efficient distribution of exam session’s time). The results can be useful for students (as a motivation to put more efforts into studying some disciplines), and for tutors and deputy deans for academic affairs (to search for “problem” students and decrease possible academic failures early).