ft_linear_regression - Predicting Car Prices with Linear Regression

📈 Project Overview

ft_linear_regression is a simple machine learning project from the 42 curriculum that implements a basic linear regression algorithm in Python to predict car prices based on mileage. The goal is to understand the mathematical foundation of linear regression and implement training, evaluation, and prediction from scratch without using ML libraries.

🚀 Features

Gradient Descent Implementation - Trains the model using gradient descent.
Manual Data Handling - Loads and parses CSV input data.
Model Training & Evaluation - Trains the model and evaluates performance visually.
Prediction Mode - Takes user input for mileage and outputs predicted price.
Data Visualization - Plots the regression line vs actual data using matplotlib.
Persistent Model Saving - Stores theta0 and theta1 in a file for reuse.

🧠 Core Concepts

Hypothesis function: price = theta0 + theta1 * mileage
Cost function (MSE): J = (1/2m) * sum((h(x) - y)^2)
Optimization via gradient descent:
- theta0 -= learning_rate * dJ/dtheta0
- theta1 -= learning_rate * dJ/dtheta1

📦 Installation

Prerequisites

Python 3.x
matplotlib (for visualization)

Install the dependencies:

pip install -r requirements.txt

🛠️ Usage

1. Training the Model

python train.py

Reads data.csv
Trains using gradient descent
Saves parameters to thetas.json

2. Predicting a Price

python predict.py

Prompts for mileage
Predicts price using saved thetas

3. Plotting the Regression

python plot.py

Plots training data and regression line

📁 File Structure

📂 ft_linear_regression/
├── data.csv         # Input data (mileage, price)
├── train.py         # Training script
├── predict.py       # User prediction interface
├── plot.py          # Data visualization
├── thetas.json      # Stored model parameters
├── requirements.txt # Python dependencies

📊 Example

Input CSV (`data.csv`):

km,price
10000,20000
25000,18000
...

Prediction:

Enter mileage: 30000
Estimated price: 17042.58$

🏗️ Future Enhancements

Add polynomial regression for better fitting.
Implement normalization for better convergence.
Add unit tests for robustness.

🏆 Credits

Developer: zekmaro
Project: Part of the 42 School curriculum
Inspiration: Andrew Ng’s ML course & scikit-learn's simplicity

🤖 A foundational step into machine learning. Explore, tweak, and enjoy learning!

zekmaro/ft_linear_regression

ErrorLooks like something went wrong!

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Languages

Error
Looks like something went wrong!

Packages