This section documents the available models. For quick iteration during development, the modules containing the models can be executed directly to test the models.


This model is based on the Long Short Term Memory deep learning model, and is implemented using Keras.


Simple ARIMA model implemented using the PyFlux library.


Simple GAS model implemented using the PyFlux library.

Vector auto regression

The VAR model is implemented following the tutorial on PyFlux.


This model uses the R library forecast <>. So in order to run it you must have R with the forecast package installed.

Currently the model adjusted to the curves is a simple arima, but more sophisticated models are available in R which can be tried.

Also this model can serve as a template for the implementation of other R based predictive models, using rpy2 as a bridge between Python and R.