FACTS-MAS: A Multi-Agent Architecture for Adaptive Forecasting from Heterogeneous Evidence
Abstract
Real-world forecasting often requires integrating heterogeneous evidence whose predictive value varies across horizons, operating regimes, and domains. We introduce FACTS-MAS, a domain-agnostic multi-agent forecasting framework that decomposes evidence into specialized channels for temporal dynamics, exogenous context, seasonality, structural information, and discrete events. Each agent generates an independently testable forecast through a common interface, while leakage-safe historical validation measures its predictive skill relative to a naive reference. These skill estimates are used to weight and combine agent forecasts, so that uninformative evidence is either removed or downweighted rather than forced to participate. A calibration procedure selects domain-specific constant means based on historical validation folds. We evaluate FACTS-MAS on weekly for-sale housing inventory from Zillow and weekly unemployment insurance claims from FRED, for 15 U.S. metropolitan areas at 4-, 8-, and 13-week forecasting horizons. Our method outperforms both a naive last-value forecast method and the TimesFM foundation model. However, a factor-adjusted exponential smoothing model performs as well as the fused forecast. The evidence therefore supports FACTS-MAS as a systematic protocol for testing, calibrating, and recomposing evidence channels before assembling them into the strongest single model for a specific application domain. Code, data, and results are available at https://github.com/Madumita24/FACTS-MAS.
Related
- Does Joint-Embedding Predictive Architecture Pretraining Help Time Series Forecasting?
- Data-Adaptive Multi-Bin Rarefying for Alpha Diversity Association Analysis under Heterogeneous Library Sizes
- Combining Weather Forecast Aggregation and State-Space Models for Adaptive Probabilistic Electricity Load Forecasting
- State-aware adaptive multimodal fusion for ultra-short-term solar irradiance forecasting using sky cloud images
- Learning Adaptive SED for heterogeneous load balancing
BibTeX
@misc{karthikeyan2026factsmas,
title = {{FACTS-MAS: A Multi-Agent Architecture for Adaptive Forecasting from Heterogeneous Evidence}},
author = {Madumita Karthikeyan and Umang Rajnikant Bid and Rong Pan},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.38225},
eprint = {2610.38225},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.38225},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - FACTS-MAS: A Multi-Agent Architecture for Adaptive Forecasting from Heterogeneous Evidence
AU - Madumita Karthikeyan
AU - Umang Rajnikant Bid
AU - Rong Pan
PY - 2026
DA - 2026/10/28
PB - arXiv
DO - 10.48550/arxiv.2610.38225
UR - https://arxiv.org/abs/2610.38225
AB - Real-world forecasting often requires integrating heterogeneous evidence whose predictive value varies across horizons, operating regimes, and domains. We introduce FACTS-MAS, a domain-agnostic multi-agent forecasting framework that decomposes evidence into specialized channels for temporal dynamics, exogenous context, seasonality, structural information, and discrete events. Each agent generates an independently testable forecast through a common interface, while leakage-safe historical validation measures its predictive skill relative to a naive reference. These skill estimates are used to weight and combine agent forecasts, so that uninformative evidence is either removed or downweighted rather than forced to participate. A calibration procedure selects domain-specific constant means based on historical validation folds. We evaluate FACTS-MAS on weekly for-sale housing inventory from Zillow and weekly unemployment insurance claims from FRED, for 15 U.S. metropolitan areas at 4-, 8-, and 13-week forecasting horizons. Our method outperforms both a naive last-value forecast method and the TimesFM foundation model. However, a factor-adjusted exponential smoothing model performs as well as the fused forecast. The evidence therefore supports FACTS-MAS as a systematic protocol for testing, calibrating, and recomposing evidence channels before assembling them into the strongest single model for a specific application domain. Code, data, and results are available at https://github.com/Madumita24/FACTS-MAS.
ER -