ddz16/TSFpaper

What it solves

This project provides a comprehensive, curated reading list of over 400 academic papers focused on time series forecasting (TSF) and spatio-temporal forecasting (STF). It helps researchers and practitioners navigate the rapidly growing literature by organizing papers by model type, conference, and specific forecasting challenges.

How it works

The repository acts as a structured index of research papers. Each entry is categorized by the forecasting type (univariate, multivariate, spatio-temporal, or irregular) and includes links to the original paper, code implementations where available, and occasional summaries. The list is continuously updated to include the latest research from top conferences and arXiv, covering emerging topics like Large Language Models (LLMs), State Space Models (Mamba), and foundation models for time series.

Who it’s for

This resource is designed for students, data scientists, and machine learning researchers who need a centralized, up-to-date reference for time series forecasting research. It is particularly useful for those looking to identify state-of-the-art models, find benchmark datasets, or understand the current landscape of deep learning applications in time series.

Highlights

• Over 400 categorized papers covering TSF and STF.
• Dedicated sections for emerging trends like LLMs, Mamba/SSM, and foundation models.
• Curated lists of popular toolkits and code libraries for practical implementation.
• Clear distinctions between forecasting types, including univariate, multivariate, and spatio-temporal methods.
• Visual markers to identify highly cited papers (🔥) and personal recommendations (🌟).