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

TL;DR

EVARS-GPR is an online Gaussian Process Regression method that detects change points and refits models to adapt to sudden shifts in seasonal data, improving forecast accuracy and computational efficiency.

Contribution

It introduces a novel online algorithm combining change point detection with data augmentation for Gaussian Process Regression in seasonal data.

Findings

20.8% lower RMSE on real-world datasets

Six-fold reduction in runtime compared to similar methods

Effective handling of sudden scale changes in seasonal data

Abstract

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 fore-casting 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 com-bines online change point detection with a refitting of the prediction model using data augmentation for samples prior to a change point. Our experiments on sim-ulated 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 con-sumption. 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 fore-casting 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.