The Complexity Research On Spline Weight Function Neural Network With High Order Composite Rational Function And Its Application

Solving the same problem can choose a variety of different algorithms. Different algorithms have different execution efficiency, so the choice of algorithm is particularly important. The main basis for the evaluation of the algorithm is the complexity of the algorithm. A good algorithm can save the time to solve the problem. Therefore, this paper has a certain practical significance to the study of complexity of the spline weight function neural network with high order composite rational function.Based on the theory of the spline weight function neural network, combining with the properties of the rational spline function, this paper construct the high order composite rational function by Hermite interpolation method. This function is used as the weight function of this neural network. Then according to the definition of the algorithm complexity, knowledge of LU methods, through calculation of the executing times of various types of operation during the algorithm executing process, we get the time complexity of the algorithm. Through MATLAB simulation experiment shows that, if the two of dimensions of the control input, output dimension, number of samples unchanged, when third variables are changed, the time consuming of the experiment is linearly related to the change of the third variables.On the basis of theoretical analysis, this paper proposes a new kind of rational spline weight function neural network to analyze the number of domestic tourism. This paper selects four primary factors that affect the domestic tourist arrivals as inputs, the number of domestic tourism as output. Then we trained and tested the network with MATLAB, the simulation results showed that the predicted results have good accuracy. Compared to the traditional BP neural network, the training speed is higher. Therefore this method can effectively predict the number of domestic tourism.