Student Performance Prediction using ML

Problem statement

Weak students are often identified only after final exams, when it is late to provide effective support.

Abstract

This project uses machine learning to predict student performance from features such as attendance, internal marks and assignment scores. The model classifies students into categories like pass, fail or grade band, helping faculty to identify at risk students early.

Components required

• Python with pandas, NumPy and scikit-learn
• CSV dataset of past student records
• Jupyter notebook or IDE
• Optional web UI using Flask or Streamlit

Block diagram

Working

The dataset is cleaned and split into training and test sets. Several models such as logistic regression and decision trees are trained and evaluated. The best model is saved. For a new student, the user enters data into a small script or web form and the model predicts the expected outcome and risk level.

Applications

• Identifying students needing academic support
• Planning remedial classes
• Education analytics for departments
• Demonstration of ML in education domain