Construct, tune, store LSTM model

[jaARke]

After data retrieval, we will need to leverage a pre-trained LSTM model to conduct forecasting.

A walkthrough guide for creating and training an LSTM model can be found here and here. We should experiment and research different layer configurations to find what achieves the best results. This is probably best done in a Jupyter notebook.

After deciding on a model, it should be stored, untrained, using Python's pickle library and made available in utils/quant/