|
Student academic success prediction system using Random Forest technique |
|---|---|
| รหัสดีโอไอ | |
| Creator | Chayaporn Kaensar |
| Title | Student academic success prediction system using Random Forest technique |
| Contributor | Wichit Sombat, Supawadee Hiranpongsin, Sirinya Baokham |
| Publisher | Faculty of Science, Ubon Ratchathani University |
| Publication Year | 2567 |
| Journal Title | Journal of Science and Science Education |
| Journal Vol. | 7 |
| Journal No. | 2 |
| Page no. | 245-259 |
| Keyword | Academic performance, Dropout rates, Predictive model, Random Forest technique, Web application |
| URL Website | https://so04.tci-thaijo.org/index.php/JSSE |
| Website title | Journal of Science and Science Education |
| ISSN | ISSN 2697-410X |
| Abstract | The challenges of academic underperformance and rising dropout rates have contributed to declining student retention rates, resulting in implications such as wasted time, increased expenses, and missed job opportunities. Consequently, the need to analyze and prepare student data is paramount. This study aims to address these challenges by: 1) investigating and constructing a predictive model of student academic performance within the Data Science and Software Innovation (DSSI), Information Technology and Communications (ICT), Chemistry, and Biology programs at the Faculty of Science, Ubon Ratchathani University, utilizing the Random Forest technique, and 2) creating a prototype web application to apply the model and summarize the findings, developing with Django and PostgreSQL. The dataset used for model construction encompasses academic records from both secondary school and university for 1,336 students enrolled in the programs spanning academic years 2017-2021. This dataset includes 11 factors: educational programs, first-term university GPA, and academic performance in 8 core subject groups: Thai Language, Mathematics, Science, Social Studies, Religion and Culture, Arts, Career and Technology, Foreign Languages, alongside high school GPA. The results of the prototype model research for DSSI, ICT, Chemistry, and Biology programs demonstrates the ability to predict student learning outcomes, with F1-measure values of 88.02%, 86.18%, 84.04% and 85.18% respectively. Consequently, the implementation of the developed system holds potential to benefit students, educators, and educational institutions in the design and implementation of more effective and tailored teaching methodologies. |
