New rules for deriving formal models from text

Abstract

Automatically converting text into formal models such as Enhanced Entity Relationship Diagrams, Class Diagrams and Relational Database Schemas is a time and effort saving process, but it is not a trivial process. There are a good number of rules required to convert Natural Language text into formal models. The purpose of this paper is to review existing rules, explain their limitations, and derive new rules for tackling those limitations and thereby helping to obtain an accurate conversion. The rules are divided into three groups: rules to define entities/classes/tables; rules to define relationship between entities/classes/tables and rules to define attributes of entities/classes/tables.