DimAug-TimesFM: Dimension Augmentation for Long-Term Cloud Demand Forecasting in Few-Shot Scenarios
Source abstract
Abstract
Accurate long-term cloud demand forecasting is critical for optimizing resource procurement and cost management in cloud computing, yet it remains challenging due to dynamic demand trends, limited historical data, and the poor generalization of existing models in few-shot scenarios. This paper proposes DimAug-TimesFM, a dimension-augmented framework for long-term cloud demand forecasting, which addresses these challenges through two key innovations. First, Delivery Period Extracting identifies critical resource delivery phases by analyzing smoothed utilization trends and differencing thresholds, enabling focused modeling on periods reflecting actual demand. Second, Dimension-Augmented TimesFM enhances the pretrained TimesFM model by integrating cross-pool data via Dynamic Time Warping based similarity matching, enriching training data while mitigating distribution discrepancies. Experiments on real-world cloud resource utilization data demonstrate that DimAug-TimesFM significantly outperforms SOTA baselines (e.g., TimesFM, DLinear, PatchTST) in both short-term (16-day) and long-term (64-day and 128-day) forecasting tasks, achieving average reductions 72.9–81.7% in RMSE. DimAug-TimesFM also exhibits better robustness in scenarios where TimesFM fails, attributed to its synergistic integration of temporal feature enhancement and cross-pool data augmentation. This work provides a practical solution for few-shot cloud demand forecasting, enabling enterprises to align resource allocation with dynamic usage patterns and reduce operational costs.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Energy Load and Power Forecasting · Traffic Prediction and Management Techniques · Air Quality Monitoring and Forecasting
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Software, Beihang University, Beijing 100191, China
Jiangxi Research Institute, Beihang University, Beijing 330200, China
References (28)
Show source-provided references
1. Storment, J.R., and Fuller, M. (2023). Cloud FinOps, O’Reilly Media, Inc.
2. Ambati, P., Bashir, N., Irwin, D., Hajiesmaili, M., and Shenoy, P. (2020, January 21–24). Hedge Your Bets: Optimizing Long-Term Cloud Costs by Mixing VM Purchasing Options. Proceedings of the 2020 IEEE International Conference on Cloud Engineering (IC2E), Sydney, NSW, Australia.10.1109/IC2E48712.2020.00018 ↗
3. Tay, C.K.L., and Shim, K.J. (2018, January 10–13). A Cloud-Based Data Gathering and Processing System for Intelligent Demand Forecasting. Proceedings of the 2018 IEEE International Conference on Big Data (Big Data), Seattle, WA, USA.10.1109/BigData.2018.8622546 ↗
4. Gao, J., Wang, H., and Shen, H. (2020, January 3–6). Machine Learning-Based Workload Prediction in Cloud Computing. Proceedings of the 2020 29th International Conference on Computer Communications and Networks (ICCCN), Honolulu, HI, USA.10.1109/ICCCN49398.2020.9209730 ↗
5. Non-Stationary Transformers: Exploring the Stationarity in Time Series Forecasting
6. DDN: Dual-Domain Dynamic Normalization for Non-Stationary Time Series Forecasting
7. Iwata, T., and Kumagai, A. (2020). Few-Shot Learning for Time-Series Forecasting. arXiv.
8. Tiny Time Mixers (TTMs): Fast Pre-Trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series
9. Tsoumplekas, G., Athanasiadis, C., Doukas, D.I., Chrysopoulos, A., and Mitkas, P. (2025). Few-Shot Load Forecasting Under Data Scarcity in Smart Grids: A Meta-Learning Approach. Energies, 18.10.3390/en18030742 ↗
10. Chen, M., Xu, Z., Zeng, A., and Xu, Q. (2023). Fraug: Frequency Domain Augmentation for Time Series Forecasting. arXiv.
11. Wirsing, K. (2020). Time Frequency Analysis of Wavelet and Fourier Transform. Wavelet Theory, BoD—Books on Demand.10.5772/intechopen.94521 ↗
12. Iterative Algorithm for Time Series Decomposition into Trend and Seasonality10.1134/S0001433821080028 ↗
13. Niu, Z., Yu, K., and Wu, X. (2020). LSTM-Based VAE-GAN for Time-Series Anomaly Detection. Sensors, 20.10.3390/s20133738 ↗
14. Mumuni, A., and Mumuni, F. (2024). Data Augmentation with Automated Machine Learning. arXiv.
15. Zhang, C., Cui, J., and Yang, B. (2019). Learning Optimal Data Augmentation Policies via Bayesian Optimization. arXiv.
16. Zhu, Q., Han, J., Chai, K., and Zhao, C. (2023). Time Series Analysis Based on Informer Algorithms. Symmetry, 15.10.3390/sym15040951 ↗
17. Taud, H., and Mas, J.F. (2018). Multilayer Perceptron (MLP). Geomatic Approaches for Modeling Land Change Scenarios, Springer.10.1007/978-3-319-60801-3_27 ↗
18. Fundamentals of Recurrent Neural Networks (RNN) and LSTM10.1016/j.physd.2019.132306 ↗
19. A Survey of Transformers10.1016/j.aiopen.2022.10.001 ↗
20. Autoformer: Decomposition Transformers with Auto-Correlation
21. Zhou, T., Ma, Z., Wen, Q., Wang, X., Sun, L., and Jin, R. (2022, January 17–23). Fedformer: Frequency Enhanced Decomposed Transformer. Proceedings of the International Conference on Machine Learning, Baltimore, MD, USA. PMLR.
22. Wu, S., Fei, H., Qu, L., Ji, W., and Chua, T. (2023). Next-GPT: Any-to-Any Multimodal LLM. arXiv.
23. Large Language Models Are Zero-Shot Time Series Forecasters
24. One Fits All: Power General Time Series Analysis by Pretrained LM
25. Das, A., Kong, W., Sen, R., and Zhou, Y. (2023). A Decoder-Only Foundation Model for Time-Series Forecasting. arXiv.
26. Jin, M., Wang, S., Ma, L., Chu, Z., Zhang, J.Y., Shi, X., Chen, P., Liang, Y., Li, Y., and Pan, S. (2023). Time-LLM: Time Series Forecasting by Reprogramming Large Language Models. arXiv.
27. Ansari, A.F., Stella, L., Turkmen, A.C., Zhang, X., Mercado, P., Shen, H., and Wang, B. (2023). Chronos: Learning the Language of Time Series. Trans. Mach. Learn. Res.
28. Liang, Y., Wen, H., Nie, Y., Jiang, Y., Jin, M., Song, D., Pan, S., and Wen, Q. (2024, January 25–29). Foundation Models for Time Series Analysis: A Tutorial and Survey. Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Barcelona, Spain.10.1145/3637528.3671451 ↗