From ER to Ontology with Natural Language Text Generation

0
94

Authors: Clare Atkins, Csaba Veres, Jennifer Sampson

Tags: 2006, conceptual modeling

We describe the automation of a novel technique (NaLER) which was originally designed to facilitate legacy database model validation. The NaLER technique uses natural language sentences built from live database content to elicit validation judgments from domain experts. However, during implementation we discovered that the method we had adopted for the automation had a serendipitous side effect in that the legacy model first had to be mapped to an upper ontology. This normally difficult process was significantly eased by the sentence templates which are defined as part of the NaLER technique. It is this novel process of mapping, and the choice of ontology it entails, which forms the focus of the paper. We therefore describe here the process of mapping to the upper model, and investigate how the motivation for modeling impacted on the choice of modeling language. Finally we describe the prototype of a tool and how it fits with the development methodology.

Read the full paper here: https://www.emmsad.org/archive/2006