AI-Powered Pricing for Used Motorcycles
Pricing used bikes is challenging due to multiple influencing factors such as brand, model, year, mileage, and condition. This system applies data-driven machine learning predictions to enhance transparency and accuracy, helping both buyers and sellers make informed pricing decisions.
After evaluating multiple models, Gradient Boosting and Random Forest delivered the best accuracy.
| Feature | Description |
|---|---|
| 🖥️ User-Friendly Web Interface | Modern and responsive design |
| 📊 Data-Driven Price Estimation | ML models trained on real-world data |
| 🔄 Dynamic Input Form | Bike brands & models loaded from a curated dataset |
| 🔔 Custom Alert System | Animated notifications for predictions |
| 🧹 Advanced Data Processing | Handles missing values, outliers, and scaling |
| 🤖 Model Comparison | Decision Tree, Random Forest, XGBoost, Gradient Boosting |
| ⚡ Optimized Performance | Hyperparameter tuning for accuracy |
Pre-Owned-Bike-Price-Prediction/
├── index.html # Landing page
├── prediction.html # Bike price prediction form
├── styles.css # Stylesheet
├── script.js # Frontend logic & form handling
├── bike.jpg # Banner image
└── README.md
Machine Learning
- Python (Pandas, NumPy, Scikit-Learn)
- Models: Decision Tree, Random Forest, XGBoost, Gradient Boosting
Web Technologies
- HTML5, CSS3, JavaScript (jQuery)
- Flask (backend API)
| Model | Performance |
|---|---|
| 🥇 Gradient Boosting | Highest Accuracy |
| 🥇 Random Forest | Highest Accuracy |
| Decision Tree | Moderate |
| XGBoost | Good |
- Python 3.x
- pip
-
Clone the repository
git clone https://github.com/ReaduanulFaridFahim/Pre-Owned-Bike-Price-Prediction.git cd Pre-Owned-Bike-Price-Prediction -
Install Python dependencies
pip install pandas numpy scikit-learn flask xgboost
-
Run the Flask server
python app.py
-
Open in browser
http://localhost:5000
Supports inputs for:
- Brand – Honda, Yamaha, Suzuki, Kawasaki, Ducati, BMW, Harley-Davidson
- Model – Dynamically loaded per brand
- Year – Manufacturing year
- Mileage – Total kilometres ridden
- Condition – Excellent / Good / Fair / Poor
- 🔹 Expanding dataset for better predictions
- 🔹 Deep Learning integration for improved accuracy
- 🔹 Price trend analysis for depreciation insights
- 🔹 Location-based pricing adjustments
This project is licensed under the MIT License.
Contributions are welcome! Please open an issue or submit a pull request.
This project bridges machine learning and e-commerce, making used bike pricing more accurate and reliable.
#MachineLearning #AI #UsedBikes #PricePrediction #DataScience #Python #WebDevelopment