FSD-CAP: Fractional Subgraph Diffusion with Class-Aware Propagation for Graph Feature Imputation

TL;DR

FSD-CAP introduces a two-stage graph feature imputation method that leverages local structure and class-aware propagation to effectively handle extreme feature missingness, achieving near full-data accuracy.

Contribution

The paper presents FSD-CAP, a novel framework combining localized fractional diffusion and class-aware refinement for robust feature imputation in highly sparse graphs.

Findings

Achieves ~80% accuracy in node classification with 99.5% missing features.

Reaches over 91% AUC in link prediction under extreme sparsity.

Outperforms existing models on large-scale and heterophily datasets.

Abstract

Imputing missing node features in graphs is challenging, particularly under high missing rates. Existing methods based on latent representations or global diffusion often fail to produce reliable estimates, and may propagate errors across the graph. We propose FSD-CAP, a two-stage framework designed to improve imputation quality under extreme sparsity. In the first stage, a graph-distance-guided subgraph expansion localizes the diffusion process. A fractional diffusion operator adjusts propagation sharpness based on local structure. In the second stage, imputed features are refined using class-aware propagation, which incorporates pseudo-labels and neighborhood entropy to promote consistency. We evaluated FSD-CAP on multiple datasets. With $99.5\%$ of features missing across five benchmark datasets, FSD-CAP achieves average accuracies of $80.06\%$ (structural) and $81.01\%$ (uniform) in node classification, close to the $81.31\%$ achieved by a standard GCN with full features. For link prediction under the same setting, it reaches AUC scores of $91.65\%$ (structural) and $92.41\%$ (uniform), compared to $95.06\%$ for the fully observed case. Furthermore, FSD-CAP demonstrates superior performance on both large-scale and heterophily datasets when compared to other models.