/references

# references:



* FFORMA: Feature-based Forecast Model Averaging Authors: Montero-Manso, Athanasopoulos, Hyndman, Talagala (2018/2019).

* Meta-learning how to forecast time series (FFORMS) 
most directly matches your question.

* A.A. Syntetos, J.E. Boylan, and J.D. Croston (2005), On the Categorization of Demand Patterns, Journal of the Operational Research Society 
(ADI/CV²) classic inventory/demand-forecasting classifier.

* Mixture-of-Linear-Experts for Long-term Time Series Forecasting
modern learned routing approach. Ronghao Ni et al. (AISTATS 2024)

* Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts
foundation-model version of routing/selection.

Rule-Based Forecasting (1992)
        ↓
Feature-based forecasting research
        ↓
FFORMS (2018-2022)
        ↓
FFORMA (2019+)
        ↓
Modern MoE / Foundation models


* book: Forecasting: Principles and Practice (Hyndman & Athanasopoulos)




* Time-MoE (ICLR 2025) as the modern MoE direction. Xiaoming Shi et al. (ICLR 2025) Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts
* Mixture-of-Linear-Experts (MoLE) as the clearest forecasting-specific expert-selection framework. Mixture-of-Linear-Experts for Long-Term Time Series Forecasting (MoLE)
Meta-learning how to Forecast Time Series (FFORMS)