基于改进的Elman神经网络的中长期径流预报

径流中长期预报长期以来一直都是人们关注的热点研究问题。现行的径流预报方法很多,传统的有时间序列法,多元回归分析法等,这些方法虽然简单易用,但是如果预报对象提供的样本容量偏小或者因子选择不够合理,都会造成预报精度偏差过大,难于有效的指导工程应用。鉴于此,本文提出一种改进的采用局部回归的Elman神经网络方法。并应用到凤滩水库优化调度的径流预报中。结果表明,与回归分析法、BP网络相比较,该方法不仅提高了算法的效率,而且提高了预报的精度,在径流预报中具有有效性和优越性。     

基于主成分分析和贝叶斯正则化方法的神经网络年最大洪峰流量预测模型探讨

针时水文预测建模中输入因子过多而导致神经网络结构规模过大,泛化能力差的问题,利用主成分分析和贝叶斯正则化方法对神经网络进行改进,优化网络结构,从而提高泛化能力。以洮儿河流域镇西站年最大洪峰流量预测为例,研究结果表明,改进的神经网络预测方法与传统的神经网络方法相比,泛化能力有显著提高,而且网络的收敛也比较稳定,实际预测中效果良好。     

人工神经网络在色彩变换中的应用

给出了一种针对数字彩色硬拷贝的输出设备，建立了由CIE色度值到设备控制值色彩变换的神经网络方法。基于蒙版方程的颜色传递机理，由色度密度变量实现色彩变换。结果表明，该色彩变换对各种色调均有较明显的校正作用，且保持了高饱和色的饱和度。     

最小生成树相位解缠中冗余去除算法

目前，相位解缠的方法中以Goldstein提出的枝切法最为经典，它通过枝切线将残差点相互连接使得残差点电荷中和，在相位展开处理的积分中，积分路径不能穿过分支，从而限制误差的传播。但是由于枝切线连接策略的不合理，往往造成连成的枝切线过多、过长，甚至多条枝切线形成闭合区域，造成不能解缠的“死区”，特别是在残差点较多的情况，这种现象更为严重。本文在最小生成树原理的基础上，提出的枝切线冗余去除算法能简单、准确地去除枝切树中的冗余，并保证了相位解缠具有最好的效果。     

粗糙神经网络在森林火灾预警中的应用

利用神经网络和粗糙集处理不确定性问题的优势，提出一种粗糙集结合神经网络进行森林火灾预测模型。通过与传统预测模型相比较，证明了该方法的有效性。     

多模式温度集成预报

基于中国国家气象中心T213模式、德国气象局业务模式和日本气象厅业务模式2 m高温度预报, 利用神经网络方法中的BP网络建立了我国600多个站的温度集成预报系统, 该预报系统的预报时效为72 h, 间隔为3 h。通过对2004年1—5月的预报结果检验, 表明:集成的温度预报结果明显优于3个模式单独的预报结果, 72 h内预报的平均绝对误差在3 ℃以内, 并且不存在明显的系统误差, 预报达到了一定的精度, 可以为预报员提供定时、定点精细的客观温度预报参考。分区的检验结果表明:不同区域预报误差存在差别, 新疆和西藏误差比较大, 而长江流域和华南地区误差很小, 并且不同区域系统误差的情况也不相同。从总体情况看, 预报误差还存在日变化, 一般来讲, 夜间的预报误差小于白天。     

基于BP神经网络的GPS导航算法

GPS导航解算中常用最小二乘算法。随着高动态用户需求精度的不断提高，且由于线性化忽略高次项，初始值精度低以及差分后剩余或放大误差的存在。导航解精度很难满足高动态用户的需求。为此，本文基于BP神经网络的非线性逼近性能。给出了基于BP神经网络的GPS导航算法。实测数据计算结果表明该算法能够真实地反映载体运动轨迹，其导航解的精度和可靠性有明显的提高。     

Integration of transfer function model and back propagation neural network for forecasting storm sewer flow in Taipei metropolis

Typhoons and storms have often brought heavy rainfalls and induced floods that have frequently caused severe damage and loss of life in Taiwan. Our ability to predict sewer discharge and forecast floods in advance during storm seasons plays an important role in flood warning and flood hazard mitigation. In this paper, we develop an integrated model (TFMBPN) for forecasting sewer discharge that combines two traditional models: a transfer function model and a back propagation neural network. We evaluated the integrated model and the two traditional models by applying them to a sewer system of Taipei metropolis during three past typhoon events (NARI, SINLAKU, and NAKR). The performances of the models were evaluated by using predictions of a total of 6 h of sewer flow stages, and six different evaluation indices of the predictions. Finally, an overall performance index was determined to assess the overall performance of each model. Based on these evaluation indices, our analysis shows that TFMBNP yields accurate results that surpass the two traditional models. Thus, TFMBNP appears to be a promising tool for flood forecasting for the Taipei metropolis sewer system. For publication in Stochastic Environmental Research and Risk Analysis.     

ANN技术在岩溶地下水可开采量评价中的应用——以山西晋祠泉域为例

以晋祠泉域为例,分析该泉域水文地质特征。应用神经网络技术(ANN)建立泉域内县代表性的难老泉岩溶地下水位与各种补排项之间定量数学模型,对该泉域地下水可开采量进行了评价。研究结果表明,所建立的岩溶地下水位多因素神经网络模型具有较好的拟合精度,仿真程度较高,所得到的地下水可采资源量评价结果与该地区地下水开发利用实际情况较为一致。同时,还计算了不同降雨条件下地下水的可开采量,使其对地下水的开采规划更具有指导意义。     

DOWNSCALING FORECAST OF MONTHLY PRECIPITATION OVER GUANGXI BASED ON BP NEURAL NETWORK MODEL

Variables fields such as enstrophy, meridional-wind and zonal-wind variables are derived from monthly 500 hPa geopotential height anomalous fields. In this work, we select original predictors from monthly 500-hPa geopotential height anomalous fields and their variables in June of 1958 - 2001, and determine comprehensive predictors by conducting empirical orthogonal function (EOF) respectively with the original predictors. A downscaling forecast model based on the back propagation (BP) neural network is built by use of the comprehensive predictors to predict the monthly precipitation in June over Guangxi with the monthly dynamic extended range forecast products. For comparison, we also build another BP neural network model with the same predictands by using the former comprehensive predictors selected from 500-hPa geopotential height anomalous fields in May to December of 1957 - 2000 and January to April of 1958 - 2001. The two models are tested and results show that the precision of superposition of the downscaling model is better than that of the one based on former comprehensive predictors, but the prediction accuracy of the downscaling model depends on the output of monthly dynamic extended range forecast.