The prediction of index in Shanghai stock based on genetic neural network
Xijie Zang, Jianli Yu · 2011
In order to avoid shortcomings of the standard BP network algorithm, the forecast model in index of shanghai stock is constructed based on genetic neural network. Involving the advantages of GA and BP, the algorithm can simultaneously complete genetic selection within a solution space to find the optimal points. Then the BP algorithm searchs the best optimal result from those points by the direction of negative gradient. Thus it not only can avoid the BP algorithm into a local minimum and slow convergence etc, but also can overcome long-search time, slow shortcomings of the GA caused by searching optimal solution in a similar form of exhaustive. Simulation indicates that the algorithm is more accuracy than the standard BP algorithm, faster in calculation and very well in applicability.