🖊️ Handwritten Digits Classification (k-NN)

A beginner-friendly Machine Learning project that implements the k-Nearest Neighbors (k-NN) algorithm from scratch in Python,
and compares it with scikit-learn’s implementation on the classic Digits dataset (8x8 grayscale images of handwritten digits 0–9).

The model is trained on the Digits dataset (8x8 grayscale images of handwritten digits 0–9).

---

Features

• k-NN implemented from scratch (no external ML libraries).
• Comparison with scikit-learn’s KNeighborsClassifier.
• Visualization of predictions vs actual digits.
• Accuracy evaluation on test data.

---

Project Structure

├── main.py # Main Python script
├── requirements.txt # Dependencies
├── results.png # Sample visualization of predictions
└── README.md # Project description

---

Results

Here’s an example of the model’s predictions:

---

🛠️ How to Run

1. Clone the repository:

   git clone https://github.com/your-username/handwritten-digits-knn.git
   cd handwritten-digits-knn

2. Install dependencies:

pip install -r requirements.txt 3. Run the project: ```bash python main.py

---

🚀 Future Improvements

1. Experiment with different values of k to compare performance.
2. Try other distance metrics (e.g., Manhattan, Cosine).
3. Extend the project to use other datasets (like MNIST).
4. Optimize performance for larger datasets.
5. Add a simple web app or GUI to make predictions interactively.