Sunspot Time Sequences Prediction Based on Process Neural Network and Quantum Particle Swarm

Aiming at the problem that difficulty of expression of the temporal accumulation in the time series prediction using artificial neural network, a prediction method which uses the process neural network is presented. The algorithm of quantum particle swarm is designed which has double chain structure and is used to train the process neural network. The algorithm used quantum bits to construct chromosomes. For the given model of process neural network, the number of genes on a chromosome is determined by the number of weight parameters and population coding is completed. Individuals in the population are updated by new quantum rotation gate and mutated by quantum non-gate. In the algorithm, each chromosome carries double chains of genes. This method can improve the possibility of optimums, expand the traverse of solution space and accelerate optimization process for process neural network. The effectiveness of the method and training algorithm are proved by the Mackey-Glass time series prediction. The simulation result shows that the method has not only high precision and fast convergence.