摘要
根据湖泊营养化程度影响因素多,评价因素与富营养化等级间的关系复杂且是非线性的.而神经网络和模糊系统各有优点,研究者将二者结合建立模糊神经网络模型用于太湖营养化评价,当固定学习速率η大于0—1内某一值,将导致网络算法不收敛,因此文中采用自适应调整学习速率,实验表明,该模型具有较快的训练速度和较高的精度.
The degree of lake eutrophication is affected by many factors and the complicated nonlinear characteristic of the relationship between the eutrophication degree and some related factors. Because neural network and fuzzy system have their superiority,in this article they are fusing. We find that BP algorithm does not convage when learning rate is bigger than a number between 0 and 1 , so we make the network adjust learning rate . The simulation results showed the validity of this method.
出处
《曲靖师范学院学报》
2007年第3期56-59,共4页
Journal of Qujing Normal University
关键词
模糊神经网络
营养化评价
模糊规则
fuzzy neural network
assessment of eutrophication
fuzzy rule