Abstract: In this thesis, we apply artificial neural network based on non-linear combinatorial forecasting method to stream flow forecast. Using the data from Taiyangtuo flood in 1996, we make combinatorial modeling and predictions of the two construeted models by using three-layer BP network. Compared to other approaches, this method improves prediction accuracy, and the results are satisfactory.

摘要： 将基于人工神经网络的非线性组合预测方法应用于径流预测中，以1996年太阳沱的洪水资料为例，对已经建立的两个模型，采用三层的BP网络进行组合模拟预测，从多方面分析比较，证明用该种方法能够提高预测精度，结果令人满意。