←Back to Projects

ML • Major / Mini

ML-Based Student Performance Prediction

Predict student performance or risk of failure using internal marks, attendance and activity features.

Objective

To build a machine learning model that predicts student performance (pass/fail or grade) based on historical academic and behavioural data.

Inputs

  • Internal marks, quiz scores, assignment scores.
  • Attendance percentage.
  • Number of backlogs, lab performance, etc.

Methodology

  1. Collect anonymized student data.
  2. Perform feature engineering and normalization.
  3. Train models: Logistic Regression, Random Forest, XGBoost.
  4. Compare accuracy, precision, recall, F1-score.
  5. Deploy as a simple web UI where faculty can enter new data and get prediction.

Outcome

The model helps identify at-risk students early so that faculty can provide timely support and remedial classes.