Two-level Decision Support System for Medical Information System

Knowledge management subsystem is frequently considered as some superstructure above an information system which serves for managerial purposes. We intend to correct this point of view taking knowledge as at list the same important part of an information system as data. We also propose that functional decomposition of a system can be spread on knowledge as well.

The two-level artificial intelligence system (AIS) is proposed. The first level supports decision making for particular domains calculating the state of business-processes in it. First-level AIS - contain their own rule engine that solves specific for it-s domain problems. It might be an either AIS of particular department of laboratory, or AIS that supports ambulance, stationery, or administrative activity as a whole, depending on volume of knowledge and size of a clinic. The second level consolidates decisions from several domains and performs decision making on a higher level.

In addition to expert knowledge, the medical information system should use Data Mining methods that allow ex-tracting rules form the data which are gained from business-processes and previously taken decisions.

AIS-s of both levels use forward chaining production system based on Rete algorithm since «if-then» rules are ex-pressive, vivid, and universal. They are implemented on CLIPS software that is freely available and can be injected