Graph Structure Learning for Spatial-Temporal Imputation: Adapting to Node and Feature Scales

TL;DR

This paper introduces GSLI, a multi-scale graph structure learning framework that dynamically adapts to heterogeneous spatial correlations in spatial-temporal data, significantly improving missing data imputation accuracy.

Contribution

The paper proposes a novel multi-scale graph structure learning framework that captures diverse spatial relationships across nodes and features, enhancing imputation performance.

Findings

GSLI outperforms existing methods on six real datasets.

Dynamic graph learning improves imputation accuracy.

Incorporating prominence modeling enhances significance in imputation.

Abstract

Spatial-temporal data collected across different geographic locations often suffer from missing values, posing challenges to data analysis. Existing methods primarily leverage fixed spatial graphs to impute missing values, which implicitly assume that the spatial relationship is roughly the same for all features across different locations. However, they may overlook the different spatial relationships of diverse features recorded by sensors in different locations. To address this, we introduce the multi-scale Graph Structure Learning framework for spatial-temporal Imputation (GSLI) that dynamically adapts to the heterogeneous spatial correlations. Our framework encompasses node-scale graph structure learning to cater to the distinct global spatial correlations of different features, and feature-scale graph structure learning to unveil common spatial correlation across features within all stations. Integrated with prominence modeling, our framework emphasizes nodes and features with greater significance in the imputation process. Furthermore, GSLI incorporates cross-feature and cross-temporal representation learning to capture spatial-temporal dependencies. Evaluated on six real incomplete spatial-temporal datasets, GSLI showcases the improvement in data imputation.