Recurrent Neural Networks for Time Series Forecasting: Current Status and Future Directions

Hansika Hewamalage, Christoph Bergmeir, Kasun Bandara
9/2/2019
cs.LGcs.NEstat.ML

Official code: HansikaPH/time-series-forecasting (75 stars, MIT, last commit 2021-12-17).

Abstract

Recurrent Neural Networks (RNN) have become competitive forecasting methods, as most notably shown in the winning method of the recent M4 competition. However, established statistical models such as ETS and ARIMA gain their popularity not only from their high accuracy, but they are also suitable for non-expert users as they are robust, efficient, and automatic. In these areas, RNNs have still a long way to go. We present an extensive empirical study and an open-source software framework of existing RNN architectures for forecasting, that allow us to develop guidelines and best practices for their use. For example, we conclude that RNNs are capable of modelling seasonality directly if the series in the dataset possess homogeneous seasonal patterns, otherwise we recommend a deseasonalization step. Comparisons against ETS and ARIMA demonstrate that the implemented (semi-)automatic RNN models are no silver bullets, but they are competitive alternatives in many situations.

View on arXivView PDF

Code Implementations(1)

HansikaPH/time-series-forecastingOfficial100%
MIT

Recurrent Neural Network Implementations for Time Series Forecasting

7526R, ShellAug 1, 20194 years agoMIT

Similar Papers

LSTM-MSNet: Leveraging Forecasts on Sets of Related Time Series with Multiple Seasonal Patterns

87% match1 repo402019

Seeking SOTA: Time-Series Forecasting Must Adopt Taxonomy-Specific Evaluation to Dispel Illusory Gains

85% match2026

Enhancing Algorithm Performance Understanding through tsMorph: Generating Semi-Synthetic Time Series for Robust Forecasting Evaluation

84% match1 repo2023

Performance Evaluation of GraphCast for Medium-Range Weather Forecasting over Brazil

84% match2026

Seasonal Encoder-Decoder Architecture for Forecasting

84% match2022

Encoding Seasonal Climate Predictions for Demand Forecasting with Modular Neural Network

83% match2023

Cite this paper

@article{hewamalage2019recurrent,
  title  = {Recurrent Neural Networks for Time Series Forecasting: Current Status and Future Directions},
  author = {Hansika Hewamalage and Christoph Bergmeir and Kasun Bandara},
  year   = {2019},
  eprint = {1909.00590},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1909.00590}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1909.00590.svg)](https://codeofpaper.com/paper/1909.00590)

Discussion