Database and Expert Systems Integration - Information Systems Reengineering, Integration and Normalization - page 308

example, the company links expert systems and databases, or the company has
bought a new expert system and links it with their existing databases.

8.1 Using a Knowledge-Based Model to Create an Expert

Database System

To provide a solution for the reengineering and/or the integration of DBSs and ESs,
a knowledge-based model with the following properties is required (Huang 1994 ):

• Ahigherlevelsynthesismodel. The best approach to integrate DBSs and ESs
is to embrace the facilities of both DBSs and ESs technologies under one um-
brella, that is, a higher order synthesis is needed. The new model will combine
high-level information modeling features, deductive capability, active capabili-
ties (i.e., integrity constraints), and the flexibility of AI-based systems with the
efficiency, security, and distributed and concurrent access provided by DBSs.
Computer scientists have investigated the use of abstract data type concepts to
define this richer data model that includes semantic data modeling concepts and
object-oriented concepts and makes no real distinction between data and knowl-
edge (in the form of rules).
• Reengineeringcapability.ThemostfeasibleapproachtointegrateDBSsandESs
was to enhance existing systems to couple both technologies. This is due to the
concept of reengineering to save on the cost of implementation. The peer-to-peer
architecture for the DBSs and ESs integration has been seen as the easiest way to
achieve the reengineering of existing DBSs and ESs.

The above criteria can be used to create a four-tier framework as depicted in Fig. 8.1 .
In this figure, the existing systems form the lower tier. The required data from these
systems is extracted using coupling classes. The coupling classes extract, and pos-

Tier 1 Tier 2 Tier 3

Tier 4

Existing or New
Expert System

Integrated
Classes

Application
Development

Coupling Classes

Existing or New
Database System