Linear Regression from Scratch vs Scikit-learn

This project implements Linear Regression using Gradient Descent from scratch in Python and compares its performance with the built-in LinearRegression model from scikit-learn.

📌 Features

• Custom Linear Regression implementation using NumPy
• Comparison with Scikit-learn's implementation
• Evaluation using Mean Squared Error (MSE)

🚀 Installation

Clone the repository and install dependencies:

pip install -r requirements.txt

📂 Requirements

• numpy
• scikit-learn

🧑‍💻 Usage

Run the script:

python main.py

You will see output like:

Custom Model Mean Squared Error: 0.0803
Sklearn Model Mean Squared Error: 0.0797

✅ Comparison

• Custom Model: Uses manual gradient descent with adjustable learning rate and iterations.
• Scikit-learn Model: Uses closed-form solution (Normal Equation).

Both should perform similarly on this dataset, with minor differences due to noise and optimization steps.

---

✨ Built for learning purposes to understand the inner workings of Linear Regression. "# linear-regression-scratch-vs-sklearn"