Knowledge Acquisition Model for Satellite Fault Diagnosis Expert System
Lianxiang Jiang, Huawang Li, Genqing Yang, Qinrong Yang · 2009
In order to solve the bottleneck problem of building an expert system, a knowledge acquisition model of fault diagnosis expert system for satellites was presented. Firstly, a data discretization algorithm based on fuzzy sets was put forward to do discretization work for decision table. Secondly, a rule extraction algorithm was brought forth to extract productive rules from decision table. Thirdly, we take an example to demonstrate how to extract productive rules for fault diagnosis expert system for satellites. The operation parameters of a satellite's power system were collected and discretized to construct a decision table. We employed attribute reduction algorithm based on discernibility matrix to do attribute reduction and then extract productive rules using the rule extraction algorithm we presented. The comparison between the rules extracted by rough sets software Rosetta and our model demonstrated the correctness and effective of our knowledge acquisition model.