Simulation and analysis of Chinese economic development based on neural network

Macroeconomic growth simulation has great importance in economic developments and policies, but there have not been any effective methods of simulation. For this we attempt to use the method of neural network (NN) to construct the model simulating the relations of GDP, financial revenue, work force and price on industry structures based on the Chinese economy. A new algorithm based on the single parameter coordinate rotatory method is provided to train and structure the multilayer feedforward neural network model. The method has much faster constringency speed than the backpropagation algorithm. The paper also provides an approach to derive the relations among the input nodes, priorities and output nodes in the NN model based on the statistical method. Furthermore, we use the statistical method to explain the simulated results and the effectiveness of the simulation. This paper provides a new perspective of economic growth.