Link Prediction based on Tensor Decomposition for the Knowledge Graph of COVID-19 Antiviral Drug

Abstract Due to the large-scale spread of COVID-19, which has a significant impact on human health and social economy, developing effective antiviral drugs for COVID-19 is vital saving lives. Various biomedical associations, e.g., drug-virus viral protein-host protein interactions, can be used building knowledge graphs. Based these sources, reasoning algorithms predict new links between viruses. To utilize various heterogeneous we proposed fusion strategy integrate results two tensor decomposition-based models (i.e., CP-N3 ComplEx-N3). Sufficient experiments indicated that our method obtained high performance (MRR=0.2328). Compared with CP-N3, mean reciprocal rank (MRR) increased by 3.3% compared ComplEx-N3, MRR 3.5%. Meanwhile, explored relationship types, there negative correlation (PCC=0.446, P-value=2.26e-194) triples predicted edge betweenness.