Monash Time Series Forecasting ArchiveDownload PDF

21 Aug 2021, 19:33 (modified: 29 Dec 2021, 04:20)NeurIPS 2021 Datasets and Benchmarks Track (Round 2)Readers: Everyone
Keywords: global time series forecasting, forecasting archive, feature analysis, baseline evaluation
TL;DR: Our paper introduces the first official time series benchmarking repository for global and multivariate time series forecasting.
Abstract: Many businesses nowadays rely on large quantities of time series data making time series forecasting an important research area. Global forecasting models and multivariate models that are trained across sets of time series have shown huge potential in providing accurate forecasts compared with the traditional univariate forecasting models that work on isolated series. However, there are currently no comprehensive time series forecasting archives that contain datasets of time series from similar sources available for researchers to evaluate the performance of new global or multivariate forecasting algorithms over varied datasets. In this paper, we present such a comprehensive forecasting archive containing 25 publicly available time series datasets from varied domains, with different characteristics in terms of frequency, series lengths, and inclusion of missing values. We also characterise the datasets, and identify similarities and differences among them, by conducting a feature analysis. Furthermore, we present the performance of a set of standard baseline forecasting methods over all datasets across ten error metrics, for the benefit of researchers using the archive to benchmark their forecasting algorithms.
Supplementary Material: pdf
URL: All information related to our datasets repository are available at https://forecastingdata.org/. The URLs of datasets repository, individual datasets and code repository are mentioned in the paper and supplementary materials.
Contribution Process Agreement: Yes
Dataset Url: Our forecasting archive is publicly available at: https://forecastingdata.org/. The tab "Datasets" contains a table representing the summary of all datasets in our archive. This table has a column named "Download", and by clicking the corresponding links mentioned under that column, the users will be redirected to the Zenodo platform (https://zenodo.org/communities/forecasting) where we have uploaded the datasets. After that, the datasets can be simply downloaded by clicking on the "Download" button on the corresponding landing page.
Dataset Embargo: Not applicable to our work. All datasets in our archive are already publicly available.
License: The datasets (https://forecastingdata.org/) and code (https://github.com/rakshitha123/TSForecasting) are licensed under a Creative Commons Attribution 4.0 International License.
Author Statement: Yes
10 Replies

Loading