/oil_leak

Leak detection and time series forecasting

Primary LanguagePython

oil_leak

Leak detection and time series forecasting

It is still on dev stage but already completely functional for daily forecasting

Build docker image like this

docker build -t ollin18/presage:0.1 .

Run the docker image this way

docker run -it --rm \
-v $(pwd)/data:/data \
-v $(pwd)/figs:/figs \
-v $(pwd)/src:/src \
ollin18/presage:0.1 /bin/bash

Inside the container run:

To train

/src/train.py --yi "2017" --yf "2017" --mi "01" --mf "03"

You can change the parameters. They stand for initial and final year and month.

To predict

/src/predict.py --y "2017" --m "12" --d "17"

The parameters stand for year, month, day.

TODO

  • Add parameters such as epoch and batch size
  • Connect to SQL database
  • Real-time predictions
  • CUDA integration
  • Pass args through docker run