Intellectual structure of biomedical informatics reflected in scholarly events

The purpose of this paper was to analyze the intellectual structure of biomedical informatics reflected in scholarly events such as conferences, workshops, symposia, and seminars. As analysis variables, ‘call for paper topics’, ‘session titles’ and author keywords from biomedical informatics-related scholarly events, and the MeSH descriptors were combined. As analysis cases, the titles and abstracts of 12,536 papers presented at five medical informatics (MI) and six bioinformatics (BI) global scale scholarly event series during the years 1999–2008 were collected. Then, n-gram terms (MI = 6,958; BI = 5,436) from the paper corpus were extracted and the term co-occurrence network was analyzed. One hundred important topics for each medical informatics and bioinformatics were identified through the hub-authority metric, and their usage contexts were compared with the k-nearest neighbor measure. To research trends, newly popular topics by 2-year period units were observed. In the past 10 years the most important topic in MI has been “decision support”, while in BI “gene expression”. Though the two communities share several methodologies, according to our analysis, they do not use them in the same context. This evidence suggests that MI uses technologies for the improvement of productivity in clinical settings, while BI uses algorithms as its tools for scientific biological discovery. Though MI and BI are arguably separate research fields, their topics are increasingly intertwined, and the gap between the fields blurred, forming a broad informatics—namely biomedical informatics. Using scholarly events as data sources for domain analysis is the closest way to approximate the forefront of biomedical informatics.