Proposals

Medical knowledge is typically gathered in catalogs or ontologies which are engineered by human experts, for example ICD-10 or SNOMED. In this work, methods for automatic learning of important medicals concepts as clusters of certain terms will be examined. We have a large database of pseudonymized textual findings from the domain of ophtalmology which can be used in the project. The goal is to develop a machine learning pipeline for creating a generic concept catalog based on the database and to compare the results with a medical ontology like ICD-10.