arXiv:2102.05314v2 Announce Type: replace Abstract: In modern time series problems, one aims at forecasting multiple time series with possible missing and noisy values. In this paper, we introduce the Sliding Mask Method (SMM) for forecasting multiple nonnegative time series by means of nonnegative matrix completion: observed noisy values and forecast/missing values are collected into matrix form, and learning is achieved by representing its rows as a convex combination of a small number of nonnegative vectors, referred to as the archetypes. We introduce two estimates, the mask Archetypal Matr
The continuous growth of time series data across various domains necessitates more robust and efficient forecasting methods, especially when data is incomplete or noisy.
Sophisticated time series forecasting with partial observations is critical for improving predictive accuracy in complex systems, impacting operations, risk management, and resource allocation.
The proposed Sliding Mask Method (SMM) offers a new algorithmic approach to handle missing and noisy data in multi-variate time series forecasting, potentially improving model robustness and efficiency.
- · AI/ML researchers
- · Data scientists
- · Industries relying on time series data (e.g., finance, logistics, healthcare)
- · Non-negative Matrix Factorization (NMF) applications
- · Traditional forecasting methods less robust to missing data
- · Organizations with siloed, incomplete time series datasets
Improved model performance in real-world applications with imperfect time series data.
Wider adoption of non-negative matrix factorization techniques in operational forecasting and predictive maintenance.
Enhanced resilience and accuracy of AI agents and autonomous systems that rely on time series predictions for decision-making.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG