Scholarlink

The first primary goal of education is to make students’ intellectual and personal development, enhancing their academic performance, critical thinking, and preparation for professional success. This study employs data from 1100 students across private universities in Bangladesh to predict academic performance using seven machine learning models: K-Nearest Neighbor (KNN), Decision Tree, Gaussian Naive Bayes, Support Vector Machine (SVM), Logistic Regression, Random Forest, and Linear Regression. Performance categories—excellent, satisfactory, moderate, unsatisfactory, and probation—are analyzed to identify critical influencing factors. Decision Tree emerged as the most accurate model with 97.10% accuracy. Notably, social media usage and probation status exhibit significant negative correlations, while factors like SGPA, CGPA, attendance, and study hours positively influence performance. This research offers actionable insights for early intervention and personalized academic support strategies, fostering improved educational outcomes. The study’s next strategy is to use deep learning to create a novel application for performance evaluation.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Machine Learning Approach for Predicting Academic Performance Using Classification Models: Based on Study of Private Universities in Bangladesh

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Arifa Tur Rahman,
  • MD. Abdullah Ibne Aziz,
  • Maria Misty Barsa,
  • Arafat Hossain Shesher,
  • Bishwaprotap Ray

摘要

The first primary goal of education is to make students’ intellectual and personal development, enhancing their academic performance, critical thinking, and preparation for professional success. This study employs data from 1100 students across private universities in Bangladesh to predict academic performance using seven machine learning models: K-Nearest Neighbor (KNN), Decision Tree, Gaussian Naive Bayes, Support Vector Machine (SVM), Logistic Regression, Random Forest, and Linear Regression. Performance categories—excellent, satisfactory, moderate, unsatisfactory, and probation—are analyzed to identify critical influencing factors. Decision Tree emerged as the most accurate model with 97.10% accuracy. Notably, social media usage and probation status exhibit significant negative correlations, while factors like SGPA, CGPA, attendance, and study hours positively influence performance. This research offers actionable insights for early intervention and personalized academic support strategies, fostering improved educational outcomes. The study’s next strategy is to use deep learning to create a novel application for performance evaluation.