Staeffy/smart_bins

Predicting the fill levels of smart bins

The repository can be used to train a forecast model for prediciting the fill levels for smart bins. Currently, it is possible to train the following model architectures:

• LSTM

Results

The LSTM model architecture achieves the following results on the test set:

• R2 = 0.77
• MSE = 0.018
• RMSE = 0.136

Plot of the forecast:

Blue: Ground Truth

Orange: Predictions

Directory overview

.
├── README.md
├── notebooks # Notebooks for data exploration
├── predictor
│   ├── config.py
│   ├── create_dataset.py
│   ├── create_features.py
│   ├── create_folders.py
│   ├── lstm_evaluate_model.py
│   ├── lstm_model.py
│   ├── lstm_trainer.py
│   └── utils
│       └── helpers.py # Helper functions
└── requirements.txt # Needed for creating Pyton env

Virtual environment:

Installing the necessary Python packages with the requirements.txt (Python 3.7)

$ python -m pip install -r requirements.txt

Running the LSTM training

1. Step: Create folder structure for assets.

python create_folders.py

Then upload your file called Smart-Bins-Messwerte(1).xlsx to the folder raw_dataset

2. Step: Create the dataset.

python create_dataset.py

3. Step: Create features for model training.

python create_features.py

4. Step: Perform LSTM model training.

python lstm_trainer.py

5. Step: Evaluate the results.

python lstm_evaluate_model.py

Contributing

1. Create your Feature Branch (git checkout -b feature/<...>)
2. Commit your Changes (git commit -m 'Add some AmazingFeature')
3. Ensure your Changes are ready to be pushed (make push-check). If you see the following message you are ready:
4. Push to the Branch (git push)
5. Open a Pull Request