Neuron Networks Classification Algorithm Based on Bionic Pattern Recognition
Traditional pattern recognition is based on the assumption that there is no any contact among similar samples. But the objective world is generally associated and similar samples comply with the law of continuity in feature space, which is the theoretical basis of Bionic Pattern Recognition. This paper introduces the theory of BPR into neu ral networks classification, firstly according to geometric meaning of the neurons in BPNN (Back Propagation Neuron Networks) and the RBFNN (Radial Basis Function Neuron Networks) in the high dimension space, a new constructive neuron networks classification algorithm based on BPR is proposed; then High dimensional geometrical shape raised by the BPR in the feature space can be covered by constructing a new type of ANN; finally a experiment proves that the algorithm is very effective.
Bionic pattern recognition (BPR) neuron neural networks (NN) high dimensiona space (HDS) classification
Qimai Chen Haiqing Zhou Yong Tang
School of Computer Science,South China Normal University,Guangzhou,P.R.China
国际会议
南宁
英文
109-120
2009-12-04（万方平台首次上网日期，不代表论文的发表时间）