高新技术产业园区发展指标的神经网络预测

Predicting of development index of state high and novel technology industrial zones based on neural network

针对目前普遍应用的BP算法存在学习速度慢和容易陷入网络瘫痪的缺陷,建立基于单参数动态搜索算法的快速人工神经网络模型,具有更快的收敛速度和更高的训练精度.应用该模型对浙江省各高新技术产业园区的发展指标进行预测,取得理想效果.

At present, the widely used BP algorithm has the shortcomings of slow convergent speed and easily reaching the network paralysis. The thesis presents a fast neural network model based on single parameter dynamic searching algorithm, which has faster convergent speed and higher accuracy than BP algorithm. It makes predicting of the development index of state high-tech and new-tech industrial zones in Zhejiang province and proves to be good.