Predicting future links between disjoint research areas using heterogeneous bibliographic information network

Document Type

Conference Proceeding

Publication Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)


Literature-based discovery aims to discover hidden connections between previously disconnected research areas. Heterogeneous bibliographic information network (HBIN) provides a latent, semi-structured, bibliographic information model to signal the potential connections between scientific papers. This paper introduces a novel literature-based discovery method that builds meta path features from HBIN network to predict co-citation links between previously disconnected literatures. We evaluated the performance of our method in predicting future co-citation links between fish oil and Raynaud’s syndrome papers. Our experimental results showed that HBIN meta path features could predict future co-citation links between these papers with high accuracy (0.851 F-Measure; 0.845 precision; 0.857 recall), outperforming the existing document similarity algorithms such as LDA, TF-IDF, and Bibliographic Coupling.

First Page


Last Page




Publication Date