Classification Tree - UNIR

A machine learning project implementing a Decision Tree Classifier to predict university admission chances based on standardized test scores.

Overview

This project was developed for UNIR (Rioja International University - Colombia) by Nicolas Zapata Alzate. It demonstrates the application of classification trees in predicting admission outcomes using student performance metrics.

Dataset

The project uses the Admission_Predict.csv dataset containing the following features:

Feature	Description
Serial No.	Unique identifier
GRE Score	Graduate Record Examination score (0-340)
TOEFL Score	Test of English as a Foreign Language score (0-120)
University Rating	University quality rating (1-5)
SOP	Statement of Purpose strength (1-5)
LOR	Letter of Recommendation strength (1-5)
CGPA	Cumulative Grade Point Average (0-10)
Research	Research experience (0 or 1)
Chance of Admit	Admission probability (0-1)

Implementation Details

Algorithm

• Model: Decision Tree Classifier
• Criterion: Gini Impurity
• Maximum Depth: 5
• Random State: 123 (for reproducibility)

Data Preprocessing

• Missing value handling with dropna(how='all')
• Feature scaling using StandardScaler
• Train-test split: 75% training, 25% testing

Features Used

• GRE Score (column index 0)
• TOEFL Score (column index 1)
• Target: Chance of Admit (column index 2, converted to categorical)

Requirements

pandas
numpy
scikit-learn
matplotlib
statsmodels

Install dependencies:

pip install pandas numpy scikit-learn matplotlib statsmodels

Usage

Run the main script:

python main.py

Output

The script generates:

1. Console output with the decision tree structure in text format
2. A visual plot (myplot.png) showing the decision tree diagram
3. Console output with accuracy metrics and model information

Project Structure

Classification_Tree_UNIR/
├── README.md
├── main.py
└── src/
    ├── Admission_Predict.csv
    ├── myplot.png
    ├── 1.png
    └── 2.png

Author

Nicolas Zapata Alzate - UNIR - Rioja International University, Colombia

License

This project is for educational purposes as part of the UNIR curriculum.