| --- |
| license: mit |
| tags: |
| - machine-learning |
| - air-quality |
| - aqi |
| - random-forest |
| - regression |
| - scikit-learn |
| library_name: scikit-learn |
| --- |
| |
| # 🌍 AQI Predictor |
|
|
| A machine learning model that predicts the **Air Quality Index (AQI)** using pollutant measurements. |
|
|
| The model was trained using air-quality data and uses four pollutant features: |
|
|
| - PM2.5 |
| - PM10 |
| - Carbon Monoxide (CO) |
| - Nitrogen Oxides (NO) |
|
|
| The trained model is saved as a Python `.pkl` file using `scikit-learn`. |
|
|
| ## 🤖 Model |
|
|
| **Algorithm:** Random Forest Regressor |
|
|
| **Task:** AQI Regression |
|
|
| **Input Features:** |
|
|
| | Feature | Description | |
| |---|---| |
| | PM2.5 | Fine particulate matter | |
| | PM10 | Particulate matter | |
| | CO | Carbon monoxide | |
| | NO | Nitrogen oxide | |
|
|
| **Output:** |
|
|
| Predicted AQI value. |
|
|
| ## 📊 Model Performance |
|
|
| The model achieved approximately: |
|
|
| - **R²:** 0.89 |
| - **RMSE:** ~22 |
|
|
| Performance was evaluated on the project dataset. Results may vary depending on the data split and preprocessing procedure. |
|
|
| ## 🔄 Prediction Pipeline |
|
|
| ```text |
| Pollutant Measurements |
| ↓ |
| PM2.5 | PM10 | CO | NO |
| ↓ |
| Trained Random Forest Model |
| ↓ |
| Predicted AQI |
