基于小波变换的脉冲耦合神经网络预测模型及其应用

被称为第三代人工神经网络的脉冲耦合神经网络(PCNN)模型,以其耦合机制、脉冲输出两大基本特性广泛应用于图像处理等领域。文章在两大基本特性的基础上对PCNN进行了改进:增加了绝对误差控制公式;连接强度以相关系数来控制,调整阈值设置为最小误差;网络的运行机制与以往的PCNN原理有所不同;将ATrous小波变换与PCNN模型相结合,形成了组合预测模型。将改进后的组合模型用于年降雨量的预测中,以求检验模型的可行性。预测结果表明,基于小波变换的PCNN组合模型用在年降雨量预测中是可行的,并取得良好效果。为进一步深化PCNN的理论、拓宽PCNN的应用领域、解决水文水资源中的预测问题提供了新的思路和方法。

PCNN Forecasting Model Based on Wavelet Transform and Its Application

Pulse Coupled Neural Network （PCNN）, called the third generation of neural network, is widely used in image processing for its basic characteristics of coupling mechanism and achieved some results. PCNN was improved based on the above basic characteristics in the paper： correlation coefficient was used to control the bonding strength and the threshold setting was adjusted by the least error; and the operation mechanism of network was different from the PCNN used in the past. A Trous transform was combined with PCNN model to form the combination forecasting model. The improved combination model was implemented in annual rainfall forecasting to check its feasibility. With good results, it demonstrated the above forecasting model was feasible. This paper is concentrated on the improved PCNN model, expanding the application area of PCNN and provides a new forecasting theories and methods in hydrology and water resources.