BlockEcho: Retaining Long-Range Dependencies for Imputing Block-Wise Missing Data

arXiv:2402.18800v1 Announce Type: new
Abstract: Block-wise missing data poses significant challenges in real-world data imputation tasks. Compared to scattered missing data, block-wise gaps exacerbate adverse effects on subsequent analytic and machine learning tasks, as the lack of local neighboring elements significantly reduces the interpolation capability and predictive power. However, this issue has not received adequate attention. Most SOTA matrix completion methods appeared less effective, primarily due to overreliance on neighboring elements for predictions. We systematically analyze the issue and propose …

abstract arxiv block capability challenges cs.lg data dependencies effects imputation issue machine machine learning power predictive stat.ml tasks type wise world