Advertisement

The Structure of Concept Models

By on

Click to learn more about author Ronald G. Ross.

This is the third and final part of a three-part series. See part one Confessions of a Concept Modeler here, and part two Creating Concept Models here.

In creating a concept model, you are designing a special kind of business blueprint, one crucial for information architecture, business analysis, and knowledge management.

Along with business concepts, a concept model comprises a set of structural relations between and among those concepts. These elements of structure come in well-defined forms, which enable a highly disciplined approach. Elements of structure come in two basic varieties, standard relations and verb concepts.

1. Standard relations. Classification and categorization bring deeper meaning (semantics) to sets of concepts that can be reasoned over. These two elements of structure do not need to be worded, just called out (often but not always graphically).

Example: We know that all mammals are animals. By calling out that connection, anything we subsequently say about animals in general automatically also applies to mammals in particular. If not, there is a flaw in our knowledge. Reasoning for free!

2. Verb concepts. These elements of structure are based on verbs and verb phrases you’ll need to form and disambiguate sentences to express your knowledge. The wordings for these verb concepts mean exactly what you say they mean – nothing more and nothing less.

Example: Suppose you say “A person steals a car.”. That doesn’t mean the person owns the car or leases the car or drives the car or anything else. It means a person steals a car. When you use the verb steals in a sentence in connection with person and car, you’ll mean exactly steals. If this verb concept is not in your concept model but you need it to communicate something, you have a hole in your business knowledge (or at the very least, your ability to communicate that knowledge). Verb concepts are essential to the expressiveness and precision of business communications.

Is a Concept Model an Ontology?

Elements of structure provide enriched semantics, hence potential computational power. Collectively, these elements of structure permit you to call your concept model a business ontology if you like.

I mean ontology in a technology-agnostic sense – that’s why I say business ontology. For example, I’m not talking about OWL (Web Ontology Language). But yes, your concept model can provide a solid grounding for use of such tools.

Be beware that not everything called a business ontology is a concept model. To be a true concept model, terms used in natural language business communications must be defined as business authors intend their meanings.

The SBVR Connection

The relevant standard for concept models is the Object Management Group’s (OMG’s) Semantics of Business Vocabulary and Business Rules (SBVR) [1]. SBVR is grounded in predicate logic as well as coordinated with standards such as OWL. You’re on very solid ground in using an approach based on SBVR.

By the way, it’s not surprising that SBVR arose in conjunction with business rules. Business rules in the large sense (not just what you see in most vendor platforms to date) have always been about expressing knowledge. They are much closer to writing legal contracts than IT requirements. But that’s a story for another day.

References

1. Refer to the SBVR Insider section of www.BRCommunity.com for more information about SBVR. SBVR was initially published in January 2008 by the Object Management Group (OMG). Its central goal is to enable the full semantics of business communication (including business rules) to be captured, encoded, analyzed for anomalies, and transferred between machines.

Extracted from Business Knowledge Blueprints: Enabling Your Data to Speak the Language of the Business, by Ronald G. Ross, 2020Download Chapter 4 free.

Leave a Reply