摘要
:用人工神经网络方法对不同水域、不同环境因子之间非线性和不确定性的复杂关系进行学习训练并预测检验。结果表明 :人工神经网络方法在模拟和预测方面均优于传统的统计回归模型 ,在资源与环境方面的应用是可行的 ,具有较强的模拟预测能力。与传统的回归模型相比 ,人工神经网络方法不要求监测数据具有很强的规律性 ,就可用训练后的网络模型对其进行预报 ,并且预测相对误差均比回归模型预测相对误差要小 ,具有一定的实用性。两个实例的应用结果表明 :人工神经网络通过神经原作用函数的简单复合就能逼近有限子集的任意非线性函数 ,而传统的统计方法则存在着如何选择模型形式及非线性优化问题 ,表现出明显的局限性 ,并且统计模型的更新工作相当繁重。
An artificial neural network(ANN) model was developed and used in different water bodies to predict timing for environmental changes as well as for the dynamics of resources.The results show that the ANN model is superior to classical statistical models (CSM) and can be used as predictive tool for highly non linear phenomena. ANN has the properties of parallel distribution processing and is adaptive in nature. Compared with CSM, two examples proved that ANN could be trained successfully, even if the available data were insufficient and irregular, while CSM showed the limit in selecting model type and non-linear optimization. There is no doubt that ANN has broad prospective in the development of resources and management of water quality.
出处
《长江流域资源与环境》
CAS
CSSCI
CSCD
2000年第2期237-241,共5页
Resources and Environment in the Yangtze Basin
关键词
人工神经网络方法
资源
环境
预测
artificial neural networks
resources
environment
prediction
application
