Car Price Prediction and Recommendation System

Project Description

This project is a machine learning-based car price prediction system. It uses a cleaned and optimized dataset to train a regression model for predicting car prices based on user input. Additionally, the system provides recommendations for similar cars from the dataset. The project includes features such as dataset preprocessing, hyperparameter optimization, and a web interface for user interaction.

Dataset

Dataset Cardekho used in this project. You can access from Kaggle

Workflow

Dataset Cleaning

1. Removing Null Values: Missing values are handled by dropping incomplete rows or filling them with appropriate values.
2. Feature Engineering:
   • Created new features like car_age from the year column.
   • Applied one-hot encoding for categorical columns such as fuel, transmission, seller_type, and brand.
3. Scaling and Normalization:
   • Used MinMaxScaler for numerical features like mileage(km/ltr/kg) and engine.
4. Outlier Removal:
   • Implemented the IQR method to remove extreme values in the selling_price column.

Dataset Optimization

• Applied TF-IDF vectorization to extract important features from car model names.
• Removed redundant columns for improved model training performance.

Model Training

Algorithm

• Trained a Random Forest Regressor on the dataset for price prediction.

Hyperparameter Optimization

• Conducted a grid search using GridSearchCV with parameters like:
  • n_estimators
  • max_depth
  • min_samples_split
  • min_samples_leaf

Metrics

• Evaluated the model using the following metrics:
  • Mean Absolute Error (MAE)
  • Mean Squared Error (MSE)
  • R² Score
  • Explained Variance

Results

• Best Parameters: Selected based on cross-validation R² scores.
• Feature importance was visualized to understand the contribution of each feature.

Visualizations

Feature Importance

A bar chart visualizing the importance of features in predicting car prices.

Actual vs Predicted Values

Scatter plot showing the relationship between actual and predicted prices.

Learning Curve

A learning curve was plotted to show the model's performance on the training and validation sets. It highlights the relationship between the number of training samples and model accuracy.

Error Distribution

Histogram depicting the distribution of prediction errors.

Web Interface

• The project includes a Flask-based web interface.
• Users can input car details to get:
  1. Predicted price for their car.
  2. Recommendations for similar cars.

Suggestion System

• A suggestion system identifies the top 5 most similar cars based on:
  • Euclidean distances between user input and dataset features.
• Provides additional insights by comparing predicted prices with actual dataset prices.

Future Improvements

• Incorporate additional features like car condition and regional pricing variations.
• Enhance the recommendation algorithm with advanced similarity metrics.
• Transition to a cloud-based architecture for real-time predictions.

Installation

To run this project locally, follow these steps:

1. Clone the Repository:

   git clone https://github.com/yourusername/car-price-prediction.git
   cd car-price-prediction

2. Install Required Dependencies:

   pip install -r requirements.txt

3. Prepare the Dataset:

   • Place the dataset file (cardekho.csv) in the root directory.

4. Train the Model:

   python train_model.py

5. Run the Web Application:

   python web.py

   Access the application at http://127.0.0.1:5000/.

Contributing

Contributions are welcome! To contribute:

1. Fork the Repository.
2. Create a Feature Branch:

   git checkout -b feature-name

3. Commit Your Changes:

   git commit -m "Describe your feature"

4. Push to Your Fork:

   git push origin feature-name

5. Open a Pull Request.