Update documentation
Browse files
README.md
CHANGED
|
@@ -1,11 +1,70 @@
|
|
| 1 |
---
|
| 2 |
-
title:
|
| 3 |
-
emoji:
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
-
sdk:
|
|
|
|
|
|
|
| 7 |
pinned: false
|
| 8 |
-
short_description: Engine performance
|
| 9 |
---
|
| 10 |
|
| 11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Engine Predictive Maintenance
|
| 3 |
+
emoji: 🔧
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: green
|
| 6 |
+
sdk: streamlit
|
| 7 |
+
sdk_version: 1.26.0
|
| 8 |
+
app_file: app.py
|
| 9 |
pinned: false
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
+
# Engine Predictive Maintenance
|
| 13 |
+
|
| 14 |
+
Predict engine failures before they happen using machine learning.
|
| 15 |
+
|
| 16 |
+
## 🎯 Features
|
| 17 |
+
- **Real-time Monitoring**: Enter sensor readings manually for instant predictions
|
| 18 |
+
- **Batch Processing**: Upload CSV files for bulk predictions
|
| 19 |
+
- **Risk Assessment**: Get risk level (Low/Medium/High) with confidence scores
|
| 20 |
+
- **Data Export**: Download predictions with timestamps for auditing
|
| 21 |
+
|
| 22 |
+
## 📊 Model Performance
|
| 23 |
+
- **Algorithm**: Random Forest with SMOTE (handles class imbalance)
|
| 24 |
+
- **Accuracy**: 92%+
|
| 25 |
+
- **F2-Score**: 0.657 (optimized for recall - prioritizes catching failures)
|
| 26 |
+
- **Training Data**: 1000+ engine sensor readings
|
| 27 |
+
|
| 28 |
+
## 🚀 Quick Start
|
| 29 |
+
|
| 30 |
+
### Manual Prediction
|
| 31 |
+
1. Enter 6 sensor readings (Lube Oil Pressure, Temperature, RPM, etc.)
|
| 32 |
+
2. Get instant failure probability and risk level
|
| 33 |
+
3. Review model confidence metrics
|
| 34 |
+
|
| 35 |
+
### Batch Prediction
|
| 36 |
+
1. Upload a CSV file with multiple engine readings
|
| 37 |
+
2. Get predictions for all rows
|
| 38 |
+
3. Download results with timestamps
|
| 39 |
+
|
| 40 |
+
## 📈 Sensor Inputs
|
| 41 |
+
- Lube Oil Pressure
|
| 42 |
+
- Lube Oil Temperature
|
| 43 |
+
- Coolant Temperature
|
| 44 |
+
- Engine RPM
|
| 45 |
+
- Fuel Pressure
|
| 46 |
+
- Coolant Pressure
|
| 47 |
+
|
| 48 |
+
## ⚙️ Technical Details
|
| 49 |
+
- **Model Format**: .joblib (scikit-learn native)
|
| 50 |
+
- **Data Processing**: Standardized features, engineered ratios
|
| 51 |
+
- **Inference**: <100ms per prediction
|
| 52 |
+
- **Deployment**: Hugging Face Spaces (auto-scaling)
|
| 53 |
+
|
| 54 |
+
## 📝 How It Works
|
| 55 |
+
1. Input sensors are standardized using training data statistics
|
| 56 |
+
2. Features are engineered (pressure ratios, temperature differences)
|
| 57 |
+
3. Random Forest classifier produces failure probability
|
| 58 |
+
4. F2-Score threshold (0.65) determines risk level classification
|
| 59 |
+
|
| 60 |
+
## 🔒 Privacy & Security
|
| 61 |
+
- All data processing happens in-browser
|
| 62 |
+
- No data is stored or logged
|
| 63 |
+
- Model weights are public and auditable
|
| 64 |
+
- Predictions are real-time only
|
| 65 |
+
|
| 66 |
+
## 🤝 Support
|
| 67 |
+
For issues or feedback, visit: https://github.com/your-org/predictive-engine-maintenance
|
| 68 |
+
|
| 69 |
+
---
|
| 70 |
+
*Built with scikit-learn, Streamlit, and Hugging Face Spaces*
|