1 December 2023 in Proceedings of the VLDB Endowment
Multiple time series forecasting plays an essential role in many applications. Solutions based on graph neural network (GNN) that deliver state-of-the-art forecasting performance use the relation graph which can capture historical correlations among time series. However, in real world, it is common that correlations among time series evolve across time, resulting in dynamic relation graph, where the future correlations may be different from those in history. To address this problem, we propose multiple time series forecasting with dynamic graph modeling (MTSF-DG) that is able to learn historical relation graphs and predicting future relation graphs to capture the dynamic correlations. We also propose a causal GNN to extract features from both kinds of relation graphs efficiently. Then we propose a reasoning network to explicitly learn the variant influence from historical timestamps to future timestamps for final forecasting. Extensive experiments on six benchmark datasets show that MTSF-DG consistently outperforms state-of-the-art baselines, and justify our design with dynamic relation graph modeling.
14 October 2025 in Nature Communications
AAAlexander de Linde Agerskov
Alexander de Linde Agerskov
+3
Understanding species-level abundance dynamics in complex microbial communities is key to managing microbial ecosystems, yet it remains a major challenge. In wastewater treatment plants (WWTPs), the presence and abundance of process-critical bacteria are essential for removing or recycling pollutants. However, individual species can fluctuate without recurring patterns. Accurately forecasting these dynamics is critical for preventing failures and guiding process optimization. We have developed a graph neural network-based model that uses only historical relative abundance data to predict future dynamics. Each model is trained and tested on individual time-series from 24 full-scale Danish WWTPs (4709 samples collected over 3–8 years, 2–5 times per month). It accurately predicts species dynamics up to 10 time points ahead (2–4 months), sometimes up to 20 (8 months). The approach, implemented as the “mc-prediction” workflow, is also tested on other datasets, including a human gut microbiome, showing its suitability for any longitudinal microbial dataset.
2 June 2025 in IEEE Transactions on Knowledge and Data Engineering
Accurate time series forecasting is crucial for optimizing resource allocation, industrial production, and urban management, particularly with the growth of cyber-physical and IoT systems. However, limited training sample availability in fields like physics and biology poses significant challenges. Existing models struggle to capture long-term dependencies and to model diverse meta-knowledge explicitly in few-shot scenarios. To address these issues, we propose MetaGP, a meta-learning-based Gaussian process latent variable model that uses a Gaussian process kernel function to capture long-term dependencies and to maintain strong correlations in time series. We also introduce Kernel Association Search (KAS) as a novel meta-learning component to explicitly model meta-knowledge, thereby enhancing both interpretability and prediction accuracy. We study MetaGP on simulated and real-world few-shot datasets, showing that it is capable of state-of-the-art prediction accuracy. We also find that MetaGP can capture long-term dependencies and can model meta-knowledge, thereby providing valuable insights into complex time series patterns.