Finding Hidden Relationships between Medical Concepts by Leveraging Metamap and Text Mining Techniques

Abstract

A lot of efforts have been made in order to make new discoveries in the biomedical filed. However, those valuable information may be hidden in text without applying appropriate text mining techniques. In this paper, I utilize MetaMap, a powerful biomedical tool provided by National Library of Medicine (NLM), along with appropriate text mining techniques, to detect hidden connections between biomedical concepts. The huge volume of Medline documents are used as data source and experimental data, where more than 20 million titles and abstracts of Medline articles are analyzed. On top of this corpus, biomedical concept queries are enabled to allow users to specify any two particular medical concepts, and the system will automatically identify potential relationships that may connect them. A graphical user interface is also developed to facilitate the search process and result presentation.