Scholarlink

<p>The success of educational institutions depends on the retention rate of their students. It is a key factor that determines whether these institutions receive accreditation, ratings, or rankings. Retention also influences the placement of students. Students are categorised as being above average, average, or below average according to academic achievement. The categorisation assists in pinpointing those with poor academic performance, and necessary training can be provided to enhance their performance. Data collection involved survey responses of 1100 students studying for Bachelor of Business Administration (BBA), B.Com, B.Pharmacy, Bachelor of Engineering (B.E), Master of Science (M.Sc), and Ph.D. programs in different Indian states with varying grading systems. Data of students’ surveys were analysed for feature importance using machine learning approaches such as random forest (RF), tuned random forest (TRF), gradient boosting, and neural network. Feature importance results of the best machine learning model, namely TRF, reveal that the age, study location, extra-curricular activities, and family income are important features when predicting the performance of students whereas other factors were less significant. Real-time data collection was performed from the departments of admissions and examinations at the university level.&#xa0;</p>

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

Predicting Student Academic Performance Using Machine Learning: A Study on Key Influential Factors and Early Risk Detection

  • 期刊论文

  • 发表日期：

    2026年7月7日

查看全文
  • Vandana Bharadi,
  • Ajay Khunteta

摘要

The success of educational institutions depends on the retention rate of their students. It is a key factor that determines whether these institutions receive accreditation, ratings, or rankings. Retention also influences the placement of students. Students are categorised as being above average, average, or below average according to academic achievement. The categorisation assists in pinpointing those with poor academic performance, and necessary training can be provided to enhance their performance. Data collection involved survey responses of 1100 students studying for Bachelor of Business Administration (BBA), B.Com, B.Pharmacy, Bachelor of Engineering (B.E), Master of Science (M.Sc), and Ph.D. programs in different Indian states with varying grading systems. Data of students’ surveys were analysed for feature importance using machine learning approaches such as random forest (RF), tuned random forest (TRF), gradient boosting, and neural network. Feature importance results of the best machine learning model, namely TRF, reveal that the age, study location, extra-curricular activities, and family income are important features when predicting the performance of students whereas other factors were less significant. Real-time data collection was performed from the departments of admissions and examinations at the university level.