Upload Mars pressure prediction model and documentation

README.md

@@ -1,93 +1,47 @@
 ---
-language: code
-tags:
-- mars
-- atmospheric-science
-- regression
-- environmental-monitoring
+title: Mars Atmospheric Pressure Predictor
+emoji: 🌡️
+colorFrom: red
+colorTo: yellow
+sdk: gradio
+sdk_version: 5.13.1
+app_file: app.py
+pinned: false
 license: mit
-metrics:
-- mae
-- rmse
-- r2_score
 ---
 
-# Mars Atmospheric Pressure Prediction Model
+# Mars Atmospheric Pressure Predictor
 
-This model predicts atmospheric pressure on Mars using environmental measurements from the REMS (Rover Environmental Monitoring Station) dataset.
+This Space hosts an interactive interface for predicting atmospheric pressure on Mars based on environmental measurements. The model was trained on data from the REMS (Rover Environmental Monitoring Station) dataset.
 
-## Model Performance
-
-- Mean Absolute Error: 34.79 Pa
-- Root Mean Squared Error: 44.81 Pa
-- R² Score: 0.397
-- Cross-validation R² Score: 0.335 (±0.102)
-- Prediction Uncertainty (±1 std): ±44.8 Pa
-
-## Feature Importance
-
-1. Maximum Ground Temperature: 0.7104 (±0.0450)
-2. Minimum Air Temperature: 0.4002 (±0.0341)
-3. Maximum Air Temperature: 0.2830 (±0.0212)
-4. Minimum Ground Temperature: 0.2138 (±0.0253)
-
-## Input Features
-
-- max_ground_temp(°C): Maximum ground temperature
-- min_ground_temp(°C): Minimum ground temperature
-- max_air_temp(°C): Maximum air temperature
-- min_air_temp(°C): Minimum air temperature
-- Seasonal features (automatically encoded)
-
-## Prediction Range
+## How to Use
 
-- Actual pressure range: 703.0 to 924.0 Pa
-- Model prediction range: 736.2 to 931.1 Pa
+1. Adjust the sliders for:
+   - Maximum Ground Temperature (°C)
+   - Minimum Ground Temperature (°C)
+   - Maximum Air Temperature (°C)
+   - Minimum Air Temperature (°C)
+   - Month (1-12)
 
-## Usage
+2. The model will predict:
+   - Atmospheric pressure in Pascals (Pa)
+   - A prediction range accounting for uncertainty
+   - The uncertainty margin (±Pa)
 
-```python
-import joblib
-
-# Load the model
-model_info = joblib.load('mars_pressure_model.joblib')
-
-# Access components
-model = model_info['model']
-scaler = model_info['scaler']
-feature_columns = model_info['feature_columns']
-
-# Make predictions (after preparing features in the same format)
-X_scaled = scaler.transform(X)
-predictions = model.predict(X_scaled)
-```
-
-## Model Details
-
-- Type: HistGradientBoostingRegressor
-- Training samples: 3,170
-- Features: 10 (including encoded seasonal features)
-- Target: Atmospheric pressure in Pascals (Pa)
-
-## Limitations
+## Model Performance
 
-- The model performs moderately with an R² score of 0.397
-- Temporal features (month, day) showed limited importance
-- Prediction uncertainty is approximately ±44.8 Pa
-- Missing values in the original dataset were handled by the model
+- Mean Absolute Error: 34.97 Pa
+- R² Score: 0.413
+- Prediction Uncertainty: ±44.2 Pa
 
-## Dataset
+## Example Values
 
-This model was trained on the REMS (Rover Environmental Monitoring Station) Mars dataset, which contains environmental measurements collected on Mars. The dataset includes temperature, pressure, and other environmental parameters measured over time.
+- Average conditions: -13°C max ground, -75°C min ground, 2°C max air, -80°C min air
+- Warmest conditions: 11°C max ground, -52°C min ground, 24°C max air, -8°C min air
+- Coldest conditions: -67°C max ground, -100°C min ground, -61°C max air, -136°C min air
 
-## Citation
+## Notes
 
-If you use this model, please cite:
-```
-@misc{mars_pressure_model_2025,
-  title={Mars Atmospheric Pressure Prediction Model},
-  author={Cline AI Assistant},
-  year={2025},
-  publisher={HuggingFace},
-  note={Based on REMS Mars Dataset}
-}
+- The model performs moderately with an R² score of 0.413
+- Predictions include uncertainty estimates
+- Temperature ranges are based on actual Mars measurements