/ES-RNN-Pytorch

This is a work in progress Pytorch implementation of the recently proposed ES-RNN by Slawek Smyl, winner of the M4 competition

Primary LanguageJupyter Notebook

ES-RNN-Pytorch

Overview

This is a work in progress Pytorch implementation of the recently proposed ES-RNN by Slawek Smyl, winner of the M4 competition.

Progress:

  • Implement Single Series Model based on Additive season and level in Pytorch
  • Implement toy datset, and Pytorch data functions
  • Test and Debug SSM on Toy Dataset
  • Implemente multiplicative season and trend
  • Implement Multi Series Model with shared Trend LSTM
  • Implement Multi Series Multi Output Model
  • Compare performance on benchmark dataset