EVARS-GPR: EVent-Triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data

Florian Haselbeck, Dominik G. Grimm

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations


Time series forecasting is a growing domain with diverse applications. However, changes of the system behavior over time due to internal or external influences are challenging. Therefore, predictions of a previously learned forecasting model might not be useful anymore. In this paper, we present EVent-triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data (EVARS-GPR), a novel online algorithm that is able to handle sudden shifts in the target variable scale of seasonal data. For this purpose, EVARS-GPR combines online change point detection with a refitting of the prediction model using data augmentation for samples prior to a change point. Our experiments on simulated data show that EVARS-GPR is applicable for a wide range of output scale changes. EVARS-GPR has on average a 20.8% lower RMSE on different real-world datasets compared to methods with a similar computational resource consumption. Furthermore, we show that our algorithm leads to a six-fold reduction of the averaged runtime in relation to all comparison partners with a periodical refitting strategy. In summary, we present a computationally efficient online forecasting algorithm for seasonal time series with changes of the target variable scale and demonstrate its functionality on simulated as well as real-world data. All code is publicly available on GitHub: https://github.com/grimmlab/evars-gpr.

Original languageEnglish
Title of host publicationKI 2021
Subtitle of host publicationAdvances in Artificial Intelligence - 44th German Conference on AI, Proceedings
EditorsStefan Edelkamp, Ralf Möller, Elmar Rueckert
PublisherSpringer Science and Business Media Deutschland GmbH
Number of pages23
ISBN (Print)9783030876258
StatePublished - 2021
Event44th German Conference on Artificial Intelligence, KI 2021 - Virtual, Online
Duration: 27 Sep 20211 Oct 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12873 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference44th German Conference on Artificial Intelligence, KI 2021
CityVirtual, Online


  • Change point detection
  • Data augmentation
  • Gaussian process regression
  • Online time series forecasting
  • Seasonal time series


Dive into the research topics of 'EVARS-GPR: EVent-Triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data'. Together they form a unique fingerprint.

Cite this