CBNU at TREC 2016 Clinical Decision Support Track

This paper describes the participation of the CBNU team at the TREC Clinical Decision Support track 2016. We propose construction of disease-centered document clusters and semantic word vectors using word embeddings. Hierarchical disease-centered document clusters are constructed based on clinical causal relationships such as disease-symptom, disease-test, and disease-treatment relationships. Semantic word vectors for medical terms are constructed by using word2vec. Documents are retrieved by expanding disease terms and semantic words for a clinical query, and by re-ranking using disease document clusters.