IDL Spring 2022 Project - Probabilistic Forecasting of stock prices with Deep AR.

Get started

1. Modify Parent Folder path in src/constants.py to current directory.

Scripts for data preparation

1. company_data.py - To process data and generate windowed time series.
2. company_random_groups.py - To process data and generate windowed time series for grouped vector inputs.

Note: The appropriate data generation scripts are executed by the appropriate notebooks below

Link to respective notebooks

1. DeepAR Baseline
2. Higher Autoregressive Context
3. Deeper Horizon Weighted Loss
4. Grouped Vector Input

Note: There are other notebooks for experiments that we have performed that we have not included in our paper.