Automated Creation of a Database for the Analysis of the Completeness of German Natural Language Requirements

The domain-independent and automated quality inspection of software requirements specifications performed by ReqInspector heavily relies on verb-subcategorisation extraction and a reference database containing a large number of verb-subcategorisation data. This current solution only considers complements and is therefore not perfectly suited for the use in ReqInspector. This thesis proposes a new extractor that improves the lexical data used by ReqInspector and evaluates the new approach to the existing one.