Automatic Relational Schema Extraction from Natural Language Requirements Specification Text
S K Geetha, G. S. Anandha Mala · 2014
2 Abstract: Recent researches have been focussed greatly on automating the extraction of information from Natural language text and this requires a large amount of domain knowledge. This paper attempts to construct a schema that depicts the interrelationship between tables extracted from natural language requirements specification text. Initially the domain knowledge is extracted from requirements specification text. Following this the schema generator extracts the attributes from the class diagram generated by the domain knowledge extractor. By prioritizing the preferences of the attributes and adjectives present in the text, the key attributes (primary and foreign) are elicited using hand crafted rules. This is done in order to derive the schema for the tables present in the requirement specification text. The extracted schemas are then represented in XML and they are stored in a relational database. Finally the database constructed can be queried to acquire the domain knowledge which eases the job of software development.