径流中长期预报的粗集模糊推理-神经网络模型

针对传统径流中长期预报模糊推理法存在的问题,基于粗集理论、模糊推理和神经网络技术,提出了基于粗集模糊推理-神经网络的径流中长期预报模型.通过比较预报因子相关关系系数值大小来选择径流的预报因子;用粗集理论约简径流中长期预报决策表,产生模糊推理规则;用BP神经网络技术学习模糊推理中的有关参数.实例验证,该模型比较客观地确定了预报因子模糊子集的隶属函数的参数,简单可行.

Middle-long Period Forecast of Runoff Based on the Model of Rough Sets Fuzzy Logic-neural Network

Aiming at the problem of fuzzy logic method for the traditional middle-long period forecast of runoff, Based on the technology of fuzzy sets, fuzzy logic-neural network, and the model is proposed that forecasts of runoff based on the fuzzy logic-neural network. This model chooses the prediction factor of the runoff, using the method of comparison with their sizes of relevance relation factors, and then rough sets is used to reduce the meter of forecast of runoff,and the fuzzy logic rules are came into being,and finally the relevant parameters in fuzzy logic with BP neural network technology are studied. The sample testifys that the relevant factors subject to function of prediction factor fuzzy sets can be ascertained objectively. The method is simple and feasible.