Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2507.21155 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918115641655296 |
|---|---|
| author | Wolff, Malcolm Li, Matthew Selvam, Ravi Kiran Zhu, Hanjing Olivares, Kin G. Ma, Ruijun Katoch, Abhinav Ramasubramanian, Shankar Cao, Mengfei Bandarra, Roberto Gopalsamy, Rahul La Vattiata, Stefania Yang, Sitan Mahoney, Michael W. |
| author_facet | Wolff, Malcolm Li, Matthew Selvam, Ravi Kiran Zhu, Hanjing Olivares, Kin G. Ma, Ruijun Katoch, Abhinav Ramasubramanian, Shankar Cao, Mengfei Bandarra, Roberto Gopalsamy, Rahul La Vattiata, Stefania Yang, Sitan Mahoney, Michael W. |
| contents | Despite significant advancements in time series forecasting, accurate modeling of time series with strong heterogeneity in magnitude and/or sparsity patterns remains challenging for state-of-the-art deep learning architectures. We identify several factors that lead existing models to systematically underperform on low-magnitude and sparse time series, including loss functions with implicit biases toward high-magnitude series, training-time sampling methods, and limitations of time series encoding methods.
SPADE-S is a robust forecasting architecture that significantly reduces magnitude- and sparsity-based systematic biases and improves overall prediction accuracy. Empirical results demonstrate that SPADE-S outperforms existing state-of-the-art approaches across a diverse set of use cases in demand forecasting. In particular, we show that, depending on the quantile forecast and magnitude of the series, SPADE-S can improve forecast accuracy by up to 15%. This results in P90 overall forecast accuracy gains of 2.21%, 6.58%, and 4.28%, and P50 forecast accuracy gains of 0.92%, 0.77%, and 1.95%, respectively, for each of three distinct datasets, ranging from 3 million to 700 million series, from a large online retailer. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_21155 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SPADE-S: A Sparsity-Robust Foundational Forecaster Wolff, Malcolm Li, Matthew Selvam, Ravi Kiran Zhu, Hanjing Olivares, Kin G. Ma, Ruijun Katoch, Abhinav Ramasubramanian, Shankar Cao, Mengfei Bandarra, Roberto Gopalsamy, Rahul La Vattiata, Stefania Yang, Sitan Mahoney, Michael W. Machine Learning Despite significant advancements in time series forecasting, accurate modeling of time series with strong heterogeneity in magnitude and/or sparsity patterns remains challenging for state-of-the-art deep learning architectures. We identify several factors that lead existing models to systematically underperform on low-magnitude and sparse time series, including loss functions with implicit biases toward high-magnitude series, training-time sampling methods, and limitations of time series encoding methods. SPADE-S is a robust forecasting architecture that significantly reduces magnitude- and sparsity-based systematic biases and improves overall prediction accuracy. Empirical results demonstrate that SPADE-S outperforms existing state-of-the-art approaches across a diverse set of use cases in demand forecasting. In particular, we show that, depending on the quantile forecast and magnitude of the series, SPADE-S can improve forecast accuracy by up to 15%. This results in P90 overall forecast accuracy gains of 2.21%, 6.58%, and 4.28%, and P50 forecast accuracy gains of 0.92%, 0.77%, and 1.95%, respectively, for each of three distinct datasets, ranging from 3 million to 700 million series, from a large online retailer. |
| title | SPADE-S: A Sparsity-Robust Foundational Forecaster |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2507.21155 |