Open AccessAttributes selection using machine learning for analysing students’ dropping out of university: a case study
Author(s) -
Tanya Pehlivanova,
Veseliedeva
Publication year - 2021
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/1031/1/012055
Subject(s) - computer science , selection (genetic algorithm) , machine learning , artificial intelligence , classifier (uml) , software , feature selection , data mining , programming language
Many students in Bulgarian universities drop out of the university before completing their studies. Identifying students at risk of dropping out allows timely taking measures for their retention. The paper presents the results of a study conducted among students of engineering programs at Trakia University - Stara Zagora. The collected data are subjected to processing, which aims to find the most important attributes that determine the risk of dropping out of university. The processing is done with Weka open source software. Different algorithms for selecting attributes with different search methods are applied. The most appropriate attribute selection algorithm was selected after applying the BayesNet classifier to the results obtained. The indicators TP rate, Precision and F-measure were compared. When applying InfoGainAttributeEval, the highest results are obtained for the accuracy of the classification. At the next stage, it is planned to expand the study among a larger number of students from different programs and create an effective forecasting model.