I am very excited to announce that we release an initial version of AugmentTS.

It is an open-source Python package for Time Series Data Augmentation using Deep Generative Models.


Some basic features of the package are:

*Time Series Data Augmentation using Deep Generative Models

*Visualizing the Latent Space of Generative Models

*Time Series Forecasting using Deep Neural Networks


Visit the home page for more information:

https://github.com/DrSasanBarak/AugmentTS


Thanks to our genius team members, @s33ssa @Mohammad_j0 and @amirabbasasadi for such a brilliant collaboration and dedication.

This repo supports a paper by our team for improving deep learning forecast using Variational Autoencoders.

In this paper, we propose a novel data augmentation-based deep forecasting framework that enhances the forecasting accuracy in different time series data. We use Variational Autoencoder (VAE) as a deep generative model for time series data augmentation. A hybrid bi-directional Long short-term memory-Convolutional neural network (biLSTM-Conv) is trained in the augmented data and then the acquired knowledge is transferred to the original dataset.
In our evaluation of real-world time series datasets, we show that the proposed method can significantly improve the accuracy of basic deep forecasting models.
We also provide empirical evidence of the efficacy of the approach against widely accepted univariate forecasting methods.

Read more about paper:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4009937