A stock predition method based on PSO and BP hybrid algorithm

In order to solve the problem of BP neural network (such as sensitive to the initial weights, easy to sink into local minimum, slow velocity of convergence and so on) in short-term stock prediction, a new stock predition method is created. Through optimizing initial weights of BP neural network by PSO of global stochastic optimization idea can the predition model which is based on the PSO and BP be built. Morever, by means of instance analysis and comparing it with traditional BP neural network forecasting method, the results show that it not only make the velocity of convergence increase, trainning error decrease but avoid to sink into local minimum and achieve better precision of prediction as well. It has some practical value in the stock prediction.