WenjieDu/Awesome_Imputation

Provides a comprehensive benchmark (TSI-Bench) comparing imputation algorithms across 170+ datasets and integrates with PyPOTS, a unified toolbox implementing neural and statistical methods like transformers, diffusion models, and graph neural networks for partially-observed time series. Curates an annotated paper collection spanning classical approaches (MICE, chained equations) through modern deep learning architectures, with reproducible experimental configurations and preprocessing pipelines via TSDB and BenchPOTS for standardized evaluation.

411 stars.