Predictive Maintenance

The data used over one hundred machines and consisting of:

• Sensor data, such as Pressure, Voltage, Rotation, and Vibration
• Failure and Maintenance history
• Machine logs such as error types.

Two classes of models are developed:

• Anomaly detection using unsupervised machine learning algorithms: Isolation Forest and AutoEncoder For more details on anomaly detection, the readers are referred to https://github.com/jvachier/Anomaly-Detection-From-Decision-Tree-to-Generative-Model
• Prediction using supervised machine learning algorithms: Logistic Regression and Random Forest

Repository Structure

.
├── README.md 
├── poetry.lock
├── pyproject.toml
├── LICENCE
├── .gitignore
├── .github/
└── src/                   
    ├── main.py         
    ├── modules/
    │   ├── data_prepration.py
    │   ├── loading.py
    │   └── models.py
    ├── pickle_files/
    ├── data/
    └── figures/