Research and implementation of computer simulation system for neural networks

Abstract

The ability of a neural network to process information and to generate a specific pattern of electrical activity is dependent on both the intrinsic biophysical properties of the individual neurons in the network and their pattern of synaptic connectivity. Through the analysis of its biophysical and biochemical structure, and incorporation of information science into medical science, computer simulations can be used to verify the known and explore the unknown, and will certainly be very significant to biomedical fields. In this paper one computer simulation system for neural networks was designed and implemented including system architecture, equivalent circuit, mathematical modeling, data representation, etc. The system architecture was designed as pyramid frame with open, controllable and measurable properties. It was implemented in platform-independent JAVA language with friendly graphical interface. And it has been widely applied.