Repository that aims to implement the WSAE-LSTM model and replicate the results of said model as defined in "A deep learning framework for financial time series using stacked autoencoders and long-short term memory" by Wei Bao, Jun Yue, Yulei Rao (2017).

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0180944

This implementation of the WSAE-LSTM model aims to address potential issues in the implementation model as defined by Bao et al. (2017) while also simultaneously addressing issues in previous attempts to implement and replicate results of said model (i.e. mlpanda/DeepLearning_Financial).

Bao W, Yue J, Rao Y (2017). "A deep learning framework for financial time series using stacked autoencoders and long-short term memory". PLOS ONE 12(7): e0180944. https://doi.org/10.1371/journal.pone.0180944

Diagram Illustrating the WSAE-LSTM model on an abstract level:

DOI:10.6084/m9.figshare.5028110 https://figshare.com/articles/Raw_Data/5028110

This repository uses a directory structured based upon Cookiecutter Datascience.

Repository package requirements/dependencies are defined in requirements.txt for pip and/or environment.yml for Anaconda/Conda.

Repository of an existing attempt to replicate above paper in PyTorch (mlpanda/DeepLearning_Financial), checked out as a git-subrepo for reference in thesubreposdirectory. This repository, subrepos/DeepLearning_Financial, will be used as a point of reference and comparison for specific components in wsae-lstm.