UCI Student Performance Prediction

This repository contains a machine learning project that predicts whether a student will pass or fail based on their demographic, social, and academic background.

The dataset used is the Student Performance Data Set from Kaggle.

Overview

The notebook (Pass_Fail_Prediction_UCI.ipynb) explores student data and implements machine learning models to classify if a student will achieve a passing final grade.

Key Features and Feature Engineering

To improve model performance and capture deeper insights, several new features were engineered from the original dataset:

• pass (Target Variable): Binary classification based on the final grade (G3 >= 10).
• study_efficiency: Ratio of studytime to absences + 1.
• social_score: Combined score of goout (going out with friends) and freetime.
• alcohol_total: Sum of workday (Dalc) and weekend (Walc) alcohol consumption.

Models Evaluated

The dataset is preprocessed with one-hot encoding for categorical features and standard scaling for numerical features. Two classification algorithms were trained and evaluated:

1. Logistic Regression

   • Accuracy: ~89.9%
   • Shows strong precision and recall for both passing and failing students.

2. Random Forest Classifier

   • Accuracy: ~88.6%
   • Robust to non-linear relationships and utilized for sample predictions.

Usage

1. Open the Jupyter Notebook (Pass_Fail_Prediction_UCI.ipynb).
2. Ensure you have the required libraries installed:

   pip install pandas numpy scikit-learn

3. Run all cells to download the dataset via the Kaggle API, train the models, and view the evaluation metrics.
4. The notebook concludes with an example of predicting the outcome for a new, hypothetical student.

Requirements

• Python 3.x
• pandas
• numpy
• scikit-learn
• kaggle (for downloading the dataset directly)