The exponential growth of the number of medical publications makes it more and more difficult for the clinical practitioner to keep track of relevant scientific progress. Computer aided screening and analysis might help to overview and classify the content of medical publications more efficiently. The new wave of enthusiasm for machine learning algorithms has launched several new possibilities to analyze and categorize documents, ranging from sports news to hospital discharge summaries. Search algorithms for the medical literature, i.e. PubMed are based on keywords and full text search. The U.S. National Library of Medicine and National Institutes of Health have made great effort to implement semantics into their data structure 1).
Most of the concepts base on the W3C ideas and standards 2).
Concepts are identified unambiguously by a Uniform Resource Identifiers (URI). This approach allows for retrieving medical literature relating defined key words efficiently, however, it remains impossible to automatically evaluate the relevance of publications. Machine learning has enabled efficient text classification for various data sources such as tweets, customer complaints, sports news and movie reviews