Neuron identification by classification tree and particle swarm optimization

In this paper, the adaptive method, which is on the basis of hierarchical classification tree and particle swarm optimzation(PSO), is proposed to classify different neurons in the international benchmark samples. Firstly, the similarity, corresponding to the maximum and the minimum of each feature, is calculated to denote the characteristic of the neuron, and then this paper designs the rules of classification tree and utilizes PSO algorithm to optimizate crucial parameters in the classification tree. To show high performance and the effectiveness of this proposed algorithm, the successful percentage of discerning the neuron can achieve to 98.04%.