Knowledge Discovery (KD) aims at extracting knowledge from large (structured or unstructured) data sets. Subfields of KD that are of special interest for GESIS and particularly for the team Knowledge Discovery are Ontology-Based Information Extraction, Social Tagging and Semantic Annotation. Research at GESIS in these fields is mostly motivated through real-world applications in the fields of Web-based Information Systems, Information Retrieval, Digital Libraries or Recommender Systems where the extracted knowledge can be used to support both: end users and knowledge base curators.
GESIS is actively researching in the areas of KD and is applying and evaluating the outcomes of these research activities in the context of its own systems and databases. Various methods can be used to support the indexing of databases, catalogues and subject-specific guides. There is a wide range of techniques that go from social to lexical, rule-based methods and fully-automatic, statistical and machine-learning methods.
New methods for automated and/or semi-automated indexing are tested and optimized for application to social science documents in support of the intellectual indexing thus far employed by GESIS.