🧠 Decision Tree Classifier - Manual Implementation

📌 Overview

This project implements a Decision Tree Classifier from scratch in three different programming languages:

• Python (required for all groups)
• C++ (second implementation)
• Rust (third implementation)

The tree is built manually by calculating entropy and information gain for each attribute to determine the best splits. The project supports .csv and .txt datasets with configurable delimiters.

---

📚 Assignment Goals

• Understand how decision trees work in classification problems.
• Build a tree step-by-step with information gain calculations.
• Implement the algorithm without using external libraries.
• Visualize the resulting decision tree (text-based or optional image).
• Perform classification with user inputs after tree construction.

---

📁 Dataset Compatibility

• Format: .csv or .txt
• Delimiters: comma (,), dot (.), or others (configurable)
• Up to 10 columns supported
• The last column is always the target/output

Example Datasets:

• weather.csv
• contact_lenses.csv
• breast_cancer.csv

---

⚙️ Features

• Works with multiple datasets
• Step-by-step calculation of entropy and information gain
• Tree structure shown in console (text-based)
• Accepts user input for prediction
• Repeatable predictions with different inputs
• Clean and readable code, well-documented

---

🚫 Restrictions

• ❌ No external machine learning libraries (e.g. sklearn, golearn, ai4r, etc.)
• ✅ Only standard libraries allowed
• 🚫 No hardcoded datasets
• ✅ Must be able to work with any compatible dataset

---

🚀 How to Run

🐍 Python

python3 decision_tree.py <dataset_file.csv>

💻 C++

Copy
Edit
g++ -std=c++20 -o decision_tree decision_tree.cpp
./decision_tree <dataset_file.csv>

🦀 Rust

Copy
Edit
cargo run --release -- <dataset_file.csv>

🖥️ Example Output

text
Copy
Edit
[Step-by-step entropy and information gain logs]
[Text-based decision tree]
Enter input (comma-separated): sunny,hot,high,false
Prediction: no