Subtractive Clustering Based RBF Neural Network Model for Outlier Detection

Outlier detection has many important applications in the field of fraud detection, network robustness analysis and intrusion detection. Some researches have utilized the neural network to solve the problem because it has the advantage of powerful modeling ability. In this paper, we propose a RBF neural network model using subtractive clustering algorithm for selecting the hidden node centers, which can achieve faster training speed. In the meantime, the RBF network was trained with a regularization term so as to minimize the variances of the nodes in the hidden layer and perform more accurate prediction. By defining the degree of outlier, we can effectively find the abnormal data whose actual output is serious deviation from its expectation as long as the output is certainty. Experimental results on different datasets show that the proposed RBF model has higher detection rate as well as lower false positive rate comparing with the other methods, and it can be an effective solution for detecting outliers.