Mobile Phone Price Prediction 📊

A regression project to predict mid-tier smartphone prices based on their technical specifications using several machine learning algorithms, with a primary focus on XGBoost.

---

Exploratory Data Analysis (EDA)

Key insights:

• Distribution analysis of features such as RAM, Battery, CPU Core, and Rear Camera.
• Correlation heatmap to visualize relationships between numerical features.
• Features like Weight and Thickness were dropped due to low impact on pricing based on EDA results.

---

Model Development & Evaluation

Machine learning models tested:

• Random Forest Regressor
• Gradient Boosting Regressor
• XGBoost Regressor (default and tuned)

📈 Performance Summary

Model	R² Score	RMSE
Random Forest	0.3825	476.80
Gradient Boosting	0.4434	452.68
XGBoost (Default)	0.6708	348.11
XGBoost (Tuned)	0.7345	312.67

The tuned XGBoost model achieved the best performance.

---

Conclusion

• The tuned XGBoost model delivered the best performance.
• The most influential features were: RAM, Battery, CPU Core, and Rear Camera.
• The model is ready to be used for predicting mid-tier smartphone prices based on technical specifications.

---

How to Run

1. Clone this repository

      git clone https://github.com/RaymussenArthur/MobilePhonePricePrediction.git
      cd MobilePhonePricePrediction

2. Install dependencies

     pip install -r requirements.txt

3. Run the EDA notebook Navigate to the Notebooks/ folder and open EDA.ipynb to explore the data and feature insights.

4. Run the Modeling notebook In the same Notebooks/ folder, open and run Modeling.ipynb to train, evaluate, and tune the models.

5. Use the trained model The final trained model is saved as a .pkl file and can be loaded using:

    import joblib
    model = joblib.load('Results/trained_model.pkl')
    prediction = model.predict(new_data)