Skip to content

README.md

Latest commit

 

History

History
61 lines (44 loc) · 2.32 KB

README.md

File metadata and controls

61 lines (44 loc) · 2.32 KB
Raw

Student Performance Analysis

Objective

Analyze student data to predict academic success (final grades and pass/fail) and identify key influencing factors.

Dataset

395 student records with demographic (e.g., age, gender) and academic features (e.g., G1, G2 grades, study time, past failures, absences), including final grade G3 and a derived pass/fail indicator.

Methods

  • Exploratory Data Analysis (EDA): Visualize distributions and relationships.
  • Linear Regression: Predict final grade (G3) from prior scores and behaviors.
  • Logistic Regression: Predict pass/fail status.
  • Error Analysis: Inspect misclassifications to understand model weaknesses.

Workflow

  1. Data loading and cleaning
  2. Exploratory Data Analysis (EDA)
  3. Feature engineering (pass/fail label)
  4. Linear regression for final grade prediction (G3)
  5. Logistic regression for pass/fail classification
  6. Model evaluation and error analysis

Key Findings

  • Prior Performance Matters: G1 and G2 scores are the dominant predictors of G3 and pass likelihood.
  • Impact of Failures: Each past failure significantly lowers the chance of passing.
  • Limited Effect of Behaviors: Study time and absences show minimal independent effect once prior grades are considered.
  • Model Performance: Achieved ~87% accuracy on pass/fail classification, with high precision for passing students.
  • Model Weakness: Struggles with borderline/improving students not well captured by past grades.

Results Summary

  • Linear Regression (G3 prediction):

    • R² score ≈ 0.78
    • Mean Absolute Error ≈ 1.34 grades

  • Logistic Regression (Pass/Fail):

    • Accuracy ≈ 87%
    • Precision (Pass) ≈ 94%
    • Recall (Pass) ≈ 87%

Limitations

  • Behavioral features (study time, absences) have limited predictive power once prior grades are included.
  • The model struggles with borderline students whose performance is improving or declining.
  • Results are dataset-specific and may not generalize to other education systems.

Tools

Python (pandas, numpy), matplotlib, seaborn, scikit-learn

How to Run

  1. Clone the repository: git clone https://github.com/ad-546/Student-Performance-Analysis.git

  2. Install dependencies: pip install -r requirements.txt

3.Open the Jupyter notebook:

jupyter notebook student_performance_analysis.ipynb