Inference

Using existing values and relationships to deduce the implications

Often a listing of the implications will make the stakeholder reassess their statements, and could lead to a modified relationship, an updated classification or even altering the underlying list of aspects.

Theory

Following the implications

Once we have defined the key aspects, classified them and populated the main relationships we can start to work out the implications. In the picture above there are four categories of things and two relationships, for example we can see that the "Quarterly Budget" meeting is held in "Boston" and that "Brian" was the "Chairman" of the meeting. So we can work out that "Brian" must visit "Boston", if we list that as an implication and are told that "Brian" never visits "Boston" then we know something needs fixing. Is our model too simplistic, is the data wrong or have we misunderstood something? These discussions all help to refine the understanding.

Often it is the implications that reveal issues, but the situations we are discussing are complex enough to make it impossible to see all the possible inferences at a glance, so we need some kind of engine to process them for us.

Practice

Rendition

From the point of view of the user an inferred relationship is just the same as a defined one. So the way we present these should use the same types of displays as the relationships do.

Implementation

Links to this page

The following pages link to here: Artificial Intelligence, Aspect, Backward Chaining, Forward Chaining, IA Best Practice, Implementing IA, Knowledge Based Systems, Model, Ontology