Artificial Neural Networks for Prediction in 5G Networks

This is a student project where some performance parameters in 5G networks are predicted using neural networks.

The paper with more details is available here.

How to run

Run Prediction within Container

•   docker compose build
    docker compose up

• Open http://localhost:8080

Run the Dataset Generation

• start the same way as when running prediction
• docker exec -it 2022ws_artificial-neural-networks-5g-nn-backend-1 bash to get into the container shell
• poetry run python simulator/dataset_generation.py
• grab the generated csv file from the directory

Run the NN Model Training

• start the same way as running within container
• docker exec -it 2022ws_artificial-neural-networks-5g-nn-backend-1 bash to get into the container shell
• poetry run python simulator/training.py
• ray saves the results to ~/ray_results, where you can grab the my_model directory of the model which performed best.
• you can also run poetry run tensorboard --logdir ~/ray_results to spawn another webserver to look at the visualization of the results at http://localhost:6006