Abstract:
To present an application of fuzzy set theory (FST) for improving backpropagation neural network (BNN) based inductive logic programming (ILP) rule approximation. With the help of FST, the approximation of the truth values of logic programs is more reasonable, before the values are sent to the BNN for learning or for recognising. Experimental results show that the recognition accuracies are in average 83.53% and 88.39% for ILP alone and ILP&BNN, respectively. Our proposed method gives the best recognition accuracy of 90.40%