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)
Abstract
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
610
Last Page
621
DOI
10.1007/978-3-319-18032-8_48
Publication Date
1-1-2015
Recommended Citation
Sebastian, Yakub; Siew, Eu Gene; and Orimaye, Sylvester Olubolu, "Predicting future links between disjoint research areas using heterogeneous bibliographic information network" (2015). Global Population Health Faculty Publications. 20.
https://doi.org/10.1007/978-3-319-18032-8_48
https://collections.uhsp.edu/global-population-health_pubs/20