BayesianLSTM-for-Time-series-Prediction

This repository is showing how to leverage Bayesian deep learning (in this case, LSTM) to do the one next-step time-series prediction with quantifying epistemic uncertainty and aleatoric uncertainty simultaneously.

The data used here is modified from the sunspot dataset: https://www.kaggle.com/robervalt/sunspots/data#

For further reference of Bayesian Deep Learning, please refer to these great papers:

(1). http://proceedings.mlr.press/v48/gal16.html

(2). https://nips.cc/Conferences/2016/Schedule?showEvent=7250

(3). https://papers.nips.cc/paper/7141-what-uncertainties-do-we-need-in-bayesian-deep-learning-for-computer-vision

The TensorFlow version is 1.14.0.

I wish this example can be helpful. All rights reserved.