Zobraziť v EDS

Student Attrition Prediction Using Machine Learning Techniques

Uložené v:
Podrobná bibliografia
Názov: Student Attrition Prediction Using Machine Learning Techniques
Autori: Asogwa, Doris Chinedu, Asogwa, Emmanuel Chibuogu, Mbonu , Emmanuel Chinedu, Nwankpa , Joshua Makuochukwu, Belonwu , Tochukwu Sunday
Zdroj: International Journal of Computer (IJC); Vol. 49 No. 1 (2023); 16-29 ; 2307-4523
Informácie o vydavateľovi: Mohammad Nassar for Researches (MNFR)
Rok vydania: 2023
Zbierka: International Journal of Computer (IJC - Global Society of Scientific Research and Researchers, GSSRR)
Predmety: Machine learning, Predictive model, Random Forest, Random Tree algorithm, Student Attrition, Feature selection method, (Java Virtual Machine (JVM), Netbeans Integrated Software Development Environment (IDE), Weka Tool, Weka Plugin
Popis: In educational systems, students’ course enrollment is fundamental performance metrics to academic and financial sustainability. In many higher institutions today, students’ attrition rates are caused by a variety of circumstances, including demographic and personal factors such as age, gender, academic background, financial abilities, and academic degree of choice. In this study, machine learning approaches was used to develop prediction models that predicted students’ attrition rate in pursuing computer science degree, as well as students who have a high risk of dropping out before graduation. This can help higher education institutes to develop proper intervention plans to reduce attrition rates and increase the probability of student academic success. Student’s data were collected from the Federal University Lokoja (FUL), Nigeria. The data were preprocessed using existing weka machine learning libraries where the data was converted into attribute related file form (arff) and resampling techniques was used to partition the data into training set and testing set. The correlation-based feature selection was extracted and used to develop the students’ attrition model and to identify the students’ risk of dropping out. Random forest and random tree machine learning algorithms were used to predict students' attrition. The results showed that the random forest had an accuracy of 79.45%, while the random tree's accuracy was 78.09%. This is an improvement over previous results where 66.14% and 57.48% accuracy was recorded for random forest and random tree respectively. This improvement was as a result of the techniques used. It is therefore recommended that applying techniques to the classification model can improve the performance of the model.
Druh dokumentu: article in journal/newspaper
Popis súboru: application/pdf
Jazyk: English
Relation: http://ijcjournal.org/index.php/InternationalJournalOfComputer/article/view/2110/778; http://ijcjournal.org/index.php/InternationalJournalOfComputer/article/view/2110
Dostupnosť: http://ijcjournal.org/index.php/InternationalJournalOfComputer/article/view/2110
Rights: Copyright (c) 2023 Doris Chinedu Asogwa, Emmanuel Chibuogu Asogwa, Emmanuel Chinedu Mbonu , Joshua Makuochukwu Nwankpa , Tochukwu Sunday Belonwu ; https://creativecommons.org/licenses/by-nc-nd/4.0
Prístupové číslo: edsbas.388E9614
Databáza: BASE
Popis
Abstrakt:In educational systems, students’ course enrollment is fundamental performance metrics to academic and financial sustainability. In many higher institutions today, students’ attrition rates are caused by a variety of circumstances, including demographic and personal factors such as age, gender, academic background, financial abilities, and academic degree of choice. In this study, machine learning approaches was used to develop prediction models that predicted students’ attrition rate in pursuing computer science degree, as well as students who have a high risk of dropping out before graduation. This can help higher education institutes to develop proper intervention plans to reduce attrition rates and increase the probability of student academic success. Student’s data were collected from the Federal University Lokoja (FUL), Nigeria. The data were preprocessed using existing weka machine learning libraries where the data was converted into attribute related file form (arff) and resampling techniques was used to partition the data into training set and testing set. The correlation-based feature selection was extracted and used to develop the students’ attrition model and to identify the students’ risk of dropping out. Random forest and random tree machine learning algorithms were used to predict students' attrition. The results showed that the random forest had an accuracy of 79.45%, while the random tree's accuracy was 78.09%. This is an improvement over previous results where 66.14% and 57.48% accuracy was recorded for random forest and random tree respectively. This improvement was as a result of the techniques used. It is therefore recommended that applying techniques to the classification model can improve the performance of the model.