Optimizing neural networks for public opinion trends prediction
Xuelian Ye, Kongyu Yang · 2015
This paper describes the method of public opinion trends prediction based on back-propagation (BP) neural networks. This paper compares two measures which are used to optimize shortcomings of the BP neural network: genetic algorithms and simulated annealing algorithm. To improve both genetic algorithms and simulated annealing, we combine these two algorithms to optimize the BP neural network. It can not only solve the dependence on the initial sample values of the BP neural network, but also prevent its falling into local minimum. It is this dual optimization on the BP neural network that will enhance the accuracy of public opinion trends prediction significantly.