基于改进遗传BP网络在糖价预测的研究

The Research in the Sugar Price Forecast Using Improved GA BP Network

首先采用灰色理论技术对市场糖价数据进行处理,而后利用改进的实数编码遗传技术优化神经网络权阀值学习训练,使其成为简单通用、快速收敛的改进BP模型。在此基础上,将该模型应用于一维的糖价市场实例预测,结果表明,提出的RINAGA BP预测模型在以糖价形式的经济数据模拟分析上,效率和精度有明显提高。

At first this paper deals with market sugar price data according to the grey theory, then optimizes power valve value study training of neural network using the improved real code GA technology, which becomes a simply general, the rapid convergence improved the BP model. Based on this, finally this paper gives an example in the market single dimensional sugar price forecast, which indicates that RINAGA BP forecast model has the excellent efficiency and the precision in the economic data simulation analysis such as sugar price.