A Novel Natural Language Processing approach to automatically Visualize Entity-Relationship Model from Initial Software Requirements

An Entity Relationship Model is a crucial design artifact used in data modelling. An additional time and numerous resources are involved in generating the ER model from preliminary software requirements manually. In contrast, Natural Language Processing is a knowledge acquisition methodology to generate several design artifacts from preliminary software requirements automatically. However, some manual pre-processing tasks are required to perform on these textual requirements before going towards the application of Natural Language Processing techniques due to which the whole process remains semi-automatic. This research study presents a novel and fully automated Natural Language Processing approach to generate the Entity Relationship Diagram model from preliminary software requirements. Entity Relationship Diagram generator tool (ERD-Gen) is developed as a part of this research study to generate the ERD model automatically from the preliminary set of requirements. Validation of this research study is performed through two benchmark case studies. The experimental outcomes proved that our proposed NLP methodology is fully automated and does not requires any manual processing step on the textual requirements.