Missing Data Estimation in Temporal Multilayer Position-Aware Graph Neural Network (TMP-GNN)

GNNs have been proven to perform highly effectively in various node-level, edge-level, and graph-level prediction tasks several domains. Existing approaches mainly focus on static graphs. However, many graphs change over time their edge may disappear, or the node/edge attribute alter from one other. It is essential consider such evolution representation learning of nodes time-varying In this paper, we propose a Temporal Multilayer Position-Aware Graph Neural Network (TMP-GNN), node embedding approach for dynamic that incorporates interdependence temporal relations into computation. We evaluate performance TMP-GNN two different representations multilayered The assessed against most popular node-level task. Then, incorporate deep framework estimate missing data compare with corresponding competent our former experiment, baseline method. Experimental results four real-world datasets yield up 58% lower ROCAUC pair-wise classification task, 96% MAE feature estimation, particularly relatively high number mean degree connectivity.