TSFpaper

Why we included this project

For anyone researching or building forecasting models, this list is a strong starting point: more than 400 papers sorted by model type rather than dumped in one flat pile. You can follow how the field evolved from RNNs and TCNs to Transformers, state-space models, and the recent LLM-based forecasting work, with each paper tagged for the scenario it targets, univariate, multivariate, or spatio-temporal. That tagging earns its keep when you need candidate architectures for a specific problem, since traffic and weather forecasting, for example, require methods that handle the spatial dimension, and the tags surface those quickly. The maintainers also fold in fresh arXiv postings alongside conference-published work, so the list reflects current trends rather than only settled classics. It is a curated reference, not a library you can install, but as a literature map it saves real time when you are weighing which model families to evaluate.