Abstract :
Neural networks often surpass decision trees in predicting pattern classifications, but their predictions cannot be explained. This algorithm´s symbolic representations make each prediction explicit and understandable. Our approach to understanding a neural network uses symbolic rules to represent the network decision process. The algorithm, NeuroRule, extracts these rules from a neural network. The network can be interpreted by the rules which, in general, preserve network accuracy and explain the prediction process. We based NeuroRule on a standard three layer feed forward network. NeuroRule consists of four phases. First, it builds a weight decay backpropagation network so that weights reflect the importance of the network´s connections. Second, it prunes the network to remove irrelevant connections and units while maintaining the network´s predictive accuracy. Third, it discretizes the hidden unit activation values by clustering. Finally, it extracts rules from the network with discretized hidden unit activation values
Keywords :
decision theory; feedforward neural nets; knowledge acquisition; knowledge based systems; pattern classification; NeuroRule; decision trees; discretized hidden unit activation values; hidden unit activation values; network accuracy; network decision process; neural networks; pattern classifications; prediction process; predictive accuracy; rule extraction; standard three layer feed forward network; symbolic representation; symbolic rules; weight decay backpropagation network; Decision trees; Feedforward systems; Humans; Learning systems; Neural networks; Pattern classification; Predictive maintenance; Transfer functions;
