IMPROVING NEURAL NETWORK CLASSIFICATION USING FURTHER DIVISION OF RECOGNITION SPACE
Zhenxiang Chen, Haiyang Wang, Ajith Abraham, Crina Groşan, Bo Yang, Yuehui Chen, Lin Wang · 2007
Abstract. Further Division of Recognition Space (FDRS) is a novel technique used for neural network classification. Recognition space is a space that is defined to categorize data sample after sample, which is mapped by neural network learning. It is divided manually into few parts to categorize samples, which can be considered as a line segment in the traditional neural network classification. In addition, the data recognition space is divided into many partitions, which will attach to different classes automatically. Experiments results using network traffic and intrusion detection data illustrate that the method has favorable performance especially with respect to the optimization speed and classification accuracy.