CNL-ER: A Controlled Natural Language for Specifying and Verbalising Entity Relationship Models

CNLER: A Controlled Natural Language for Specifying and Verbalising Entity Relationship Models Bayzid Ashik Hossain, Gayathri Rajan and Rolf Schwitter Department of Computing Macquarie University, Sydney, Australia {bayzid-ashik.hossain|rolf.schwitter}@mq.edu.au gayathri.rajan@students.mq.edu.au Abstract The first step towards designing an informa- tion system is conceptual modelling where do- main experts and knowledge engineers iden- tify the necessary information together to build an information system. Entity relationship modelling is one of the most popular concep- tual modelling techniques that represents an information system in terms of entities, at- tributes and relationships. Entity relationship models are constructed graphically but are of- ten difficult to understand by domain experts. To overcome this problem, we suggest to ver- balise these models in a controlled natural lan- guage. In this paper, we present CNLER, a controlled natural language for specifying and verbalising entity relationship (ER) mod- els that not only solves the verbalisation prob- lem for these models but also provides the ben- efits of automatic verification and validation, and semantic round-tripping which makes the communication process transparent between the domain experts and the knowledge engi- neers. 1 Introduction An information system is a piece of software that has integrated components for organizing and ana- lyzing data to aid decision making in an organiza- tion (Laudon and Laudon, 2015). One of the ma- jor roles of an information system is to accumu- late data, turn it into information and later trans- form that information into organizational knowl- edge (Bourgeois, 2014). To be successful an infor- mation system always depends on a good design and conceptual modelling is the first step in the de- sign process (Olivé, 2007). Information systems are best specified on the conceptual level using a language with names for individuals, concepts, and relations that occur in the application domain. Such a language is easy to understand by the do- main experts and enhances correctness, compati- bility, productivity and clarity in information sys- tem design (Halpin, 1998). Conceptual modelling involves different parties (e.g., domain experts and knowledge engineers) who brainstorm together to identify the necessary information for building the system (Hossain and Schwitter, 2018). Af- ter identifying the required information, knowl- edge engineers build a conceptual model of the in- formation system by using conceptual modelling techniques such as entity relationship modelling (ERM) (Richard, 1990; Frantiska, 2018), object oriented modelling (UML) (O’Regan, 2017), or object role modelling (ORM) (Halpin, 2009). One of the problems with these models is that they are constructed graphically and as a result they are often hard to understand for domain ex- perts (Jarrar et al., 2006). Another problem with these conceptual modelling techniques is that they have no formal semantics; therefore, they are not machine comprehensible, do not support au- tomatic verification and validation nor automatic reasoning (Calvanese, 2013). To overcome these problems previous works used logic in parallel with traditional conceptual modelling techniques (Lutz, 2002; Berardi et al., 2005; Franconi et al., 2012). There are tools (Fil- lottrani et al., 2012; Lembo et al., 2016b,a) that al- low knowledge engineers to …

Content truncated to 500 words