Inference engines in expert systems perform two functions: they interpret or diagnose the database using the knowledge base, and they recommend or prescribe actions. The inference engine is in reality a set of rules that translate into "if-then" statements, such as "If A is given, then B must be the case, and for this reason C is recommended."
As noted, in ENSCAN, issues are diagnosed in accordance with their position in the magnitude/awareness (MAG) and importance/impact (IMP) matrix. Issues at the lower left-hand side of the matrix are characterized by low MAG and low IMP. These are issues that have attracted little attention from the media or awareness on the part of the issue monitor; in addition these issues must be perceived as having a low impact on the organization. The inference engine assigns one of three meanings—C, T, or O—to such issues and recommends simply continuing to identify them. No further movement along the four-step process is recommended.
Issues that fall in the upper left-hand side of the matrix are high MAG, low IMP issues. They are at the center of media attention, but managers expect them to have a low impact on the organization's present and future objectives, strategies, and plans of action. These issues also are categorized as either constraints, threats, or opportunities. For such issues the inference engine recommends that the organization continue to perform only the first two steps of the process: identification and evaluation.
The third possibility is for an issue to fall in the upper right-hand corner of the matrix, when it is judged to have both a high MAG and a high IMP. Such issues are the hottest and the most urgent ones. The inference engine interprets them as threats and recommends immediate action along the four steps of the process: the identified issues should be more carefully evaluated using a well-known technique such as cross impact analysis, trend impact analysis, or the Delphi method and then incorporated and translated.
Finally, in the lower right-hand side of the matrix are issues that have a high IMP and low MAG. These issues are interpreted by the inference engine as opportunities, and continuation of actions I and II—identification and evaluation—is recommended.
- 821 reads
