Obtaining E-R diagrams semi-automatically from natural language specifications

Meziane, F ORCID: https://orcid.org/0000-0001-9811-6914 and Vadera, S ORCID: https://orcid.org/0000-0001-6041-2646 2004, Obtaining E-R diagrams semi-automatically from natural language specifications , in: Sixth International Conference on Enterprise Information Systems (ICEIS 2004), 14-17 April 2004, Universidade Portucalense, Porto, Portugal.

PDF (Author version)
Download (496kB) | Preview


Since their inception, entity relationship models have played a central role in systems specification, analysis and development. They have become an important part of several development methodologies and standards such as SSADM. Obtaining entity relationship models, can however, be a lengthy and time consuming task for all but the very smallest of specifications. This paper describes a semi-automatic approach for obtaining entity relationship models from natural language specifications. The approach begins by using natural language analysis techniques to translate sentences to a meaning representation language called logical form language. The logical forms of the sentences are used as a basis for identifying the entities and relationships. Heuristics are then used to suggest suitable degrees for the identified relationships. This paper describes and illustrates the main phases of the approach and presents a summary of the results obtained when it is applied to a case study.

Item Type: Conference or Workshop Item (Poster)
Additional Information: In volume 1 of conference proceedings ('Databases and information systems integration')
Themes: Subjects / Themes > Q Science > QA Mathematics > QA075 Electronic computers. Computer science
Subjects outside of the University Themes
Schools: Schools > School of Computing, Science and Engineering > Salford Innovation Research Centre
Refereed: Yes
Related URLs:
Depositing User: Prof Farid Meziane
Date Deposited: 02 Mar 2009 14:58
Last Modified: 15 Feb 2022 17:40
URI: https://usir.salford.ac.uk/id/eprint/1805

Actions (login required)

Edit record (repository staff only) Edit record (repository staff only)