EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep learning in time series forecasting with transformer models and RNNs

Rogerio Pereira Dos Santos, João P Matos-Carvalho, Valderi R Q Leithardt iD

PeerJ Computer Science · 24.07.2025 · 11 · e3001

DOI: 10.7717/peerj-cs.3001 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PeerJ Computer Science, online erschienen am 24.07.2025. Verfasst von Rogerio Pereira Dos Santos und 2 weiteren Autor:innen. Beteiligte Einrichtung: COPELABS, Universidade Lusófona de Humanidades e Technologias, Lisboa, Portugal. Die Arbeit wurde bisher 25-mal zitiert (Crossref) und nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Given the increasing need for accurate weather forecasts, the use of neural networks, especially transformer and recurrent neural networks (RNNs), has been highlighted for their ability to capture complex patterns in time series. This study examined 14 neural network models applied to forecast weather variables, evaluated using metrics such as median absolute error (MedianAbsE), mean absolute error (MeanAbsE), maximum absolute error (MaxAbsE), root mean squared percent error (RMSPE), and root mean square error (RMSE). Transformer-based models such as Informer, iTransformer, Former, and patch time series transformer (PatchTST) stood out for their accuracy in capturing long-term patterns, with Informer showing the best performance. In contrast, RNN models such as auto-temporal convolutional networks (TCN) and bidirectional TCN (BiTCN) were better suited to short-term forecasting, despite being more prone to significant errors. Using iTransformer it was possible to achieve a MedianAbsE of 1.21, MeanAbsE of 1.24, MaxAbsE of 2.86, RMSPE de 0.66, and RMSE de 1.43. This study demonstrates the potential of neural networks, especially transformers, to improve accuracy, providing a practical and theoretical basis for selecting the most suitable models for predictive applications.

Schlagwörter

Stichwörter der Autor:innen

Accuracy in forecastingDeep learningNeural networksPredictive applicationsRecurrent neural networks (RNNs)Transformer models

Autor:innen und Institutionen

  • COPELABS, Universidade Lusófona de Humanidades e Technologias, Lisboa, Portugal.
  • LASIGE, Faculdade de Ciências, Universidade de Lisboa, Lisboa, Portugal.
  • Center of Technology and Systems (UNINOVA-CTS) and Associated Lab of Intelligent Systems (LASI), Universidade Nova de Lisboa, Lisboa, Caparica, Portugal.
  • Instituto Universitário de Lisboa (ISCTE-IUL), ISTAR, Lisboa, Portugal.

ORCID-Profile: Valderi R Q Leithardt

Zitieren

APA

Dos Santos, R. P., Matos-Carvalho, J. P., & Leithardt, V. R. Q. (2025). Deep learning in time series forecasting with transformer models and RNNs. PeerJ Computer Science, 11, e3001. https://doi.org/10.7717/peerj-cs.3001

Vancouver

Dos Santos RP, Matos-Carvalho JP, Leithardt VRQ. Deep learning in time series forecasting with transformer models and RNNs. PeerJ Computer Science. 2025;11:e3001. doi:10.7717/peerj-cs.3001

BibTeX
@article{DosSantos2025,
  author = {Dos Santos, Rogerio Pereira and Matos-Carvalho, João P and Leithardt, Valderi R Q},
  title = {Deep learning in time series forecasting with transformer models and RNNs},
  journal = {PeerJ Computer Science},
  year = {2025},
  volume = {11},
  pages = {e3001},
  doi = {10.7717/peerj-cs.3001}
}
Kostenlos starten →