This paper describes a method of rule extraction from trained artificial neural networks. The statement of the problem is given. The aim of rule extraction procedure and suitable neural networks for rule extraction are outlined. The RULEX rule extraction algorithm is discussed that is based on the radial basis function (RBF) neural network. The extracted rules can help discover and analyze the rule set hidden in data sets. The paper contains an implementation example, which is shown through standalone IRIS data set.
A discernibility matrix-based attribute reduction algorithm of decision table is introduced in this paper, which takes the importance of attributes as the heuristic message. This method solves the problem of the attribute selection when the frequencies of decision table attributes are equal. The result shows that this method can give out simple but effective method of attribute reduction.
The speed signal of engine contains abundant information. This paper introduces rough set theory for feature extraction from engine's speed signals, and proposes a method of mining useful information from a mass of data. The result shows that the discernibility matrix algorithm can be used to reduce attributes in decision table and eliminate unnecessary attributes, efficiently extracted the features for evaluating the technical condition of engine.
Designing a decompositional rule extraction algorithm for neural networks with bound decomposition tree