Abstract:

According to the characteristic of nonlinearity，uncertainty，time variation,this paper presents high-tech knowledge  innovation capacity evaluation index system，and puts forward an improved fuzzy neural network evaluation model combined  with particle swarm optimization. This model can combine multiple concurrent time varying fuzzy neural network algorithm  and realize network of learning and accurate reasoning, by  evolution  preset network connection weights, threshold and  compensation parameters with particle swarm optimization.Through simulating application, it has been proved that this  model structure and the algorithm are feasible and facilitate for computer implementation, and get the overall  convergence speed and generalization ability, convergence precision of superior original learning algorithm.