1. Student's Performance Prediction Model and Affecting Factors Using Classification Techniques
- Author
-
Hussain, Asif, Khan, Muzammil, and Ullah, Kifayat
- Abstract
Educational institutions are creating a considerable amount of data regarding students, faculty and related organs. This data is an essential asset for academic institutions as it has valuable insights, knowledge and intelligence for the policymakers. Students are the fundamental entities and primary source of data creation in any educational environment. The educational institutions need to distinguish students who are weak in their studies and require special attention and monitoring to improve their learning behaviours, future academic performances and factors that can affect their interpretation. This paper adopted a hybrid classification model using Decision tree and support vector machine (SVM) algorithms to predict students' academic performance. We statistically analyzed and identified factors that can affect students' academic performance. The dataset used is collected from Bachelor students of the City University of Science and Information Technology (CUSIT). The experimental results revealed 71.79%, 74.04% and 78.85% for decision tree, and 69.87%, 74.04% and 71.15% accuracy for SVM models respectively for different splits. The study identified seven different factors that can directly affect the students' performance associated with educational institutions and social networks. Factors like "time spent on social networks," "type of games playing on mobiles," and "time spent on playing mobile games" significantly affect students' performance.
- Published
- 2022
- Full Text
- View/download PDF