摘要
为了更有效地对水环境质量进行综合评价,论文提出了一种改进的T-S(Takagi-Sugeno)模糊神经网络水质评价模型,该模型首先通过减法聚类确定模糊C均值聚类(FCM)的初始聚类中心和聚类数目,改善传统FCM算法对聚类中心初值选取的随机性及样本的敏感性,降低陷入局部最优解的可能性。将减法聚类改进的FCM算法应用到T-S模糊神经网络的特征提取中,对T-S模糊神经网络模型进行结构辨识,提高评价模型的准确性和收敛速度。通过与传统的T-S模糊神经网络比较,水质评价结果准确率更高。
In order to evaluate the water quality more effectively, an improved T-S fuzzy neural network model was presented in this paper. The model determined the initial cluster centers and the number of clusters of FCM by subtractive clustering algorithm, and then it could reduce the randomness and sensitivity of the FCM clustering center initial selected value, and decreased the possibility of the local minimum. The improved FCM was used in the feature extraction and structure identification of the T-S fuzzy neural network to improve the accuracy and convergence rate. Compared with the traditional T-S fuzzy neural network, the water quality evaluation result worked out by the improved model was more accurate.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2013年第10期1197-1202,共6页
Computers and Applied Chemistry
基金
扬州市环境保护局资助项目(YHK0902)
江苏省科技厅软科学项目(BR2008098)
(BR2012043)
关键词
FCM
T-S模型
神经网络
减法聚类
水质综合评价
FCM
T-S model
neural network
subtractive clustering method
water quality comprehensive assessment