An Algorithm for Mining Decision Rules Based on Decision Network and Rough Set Theory

Abstract

Knowledge acquisition continues to be a challenging and time consuming task in building decision support systems. Rule induction is a data mining process for acquiring knowledge in terms of decision rules from a number of specific ’examples’ to explain the inherent causal relationship between conditional factors and a given decision. This paper introduces a method of rules extraction for fault diagnosis based on rough set theory and decision network. The fault diagnosis decision system attributes are reduced firstly, and then a decision network with different reduced levels is constructed. Initialize the network’s node with the attribute reduction sets and extract the decision rule sets according to the node of the decision network. In addition, the coverage degree and the certainty factor were applied to filter noise and evaluate the extraction rules. The proposed methodology cannot only set up rational and succinct diagnosis model for large complicated power system but also it can dispose uncertainty information of substation and get correct diagnosis result under incomplete information. At last an example is given; the result indicates that the method can diagnosis single fault and multi-fault efficiently and could be used to assist operators in their decision-making processes.