基于减法聚类产生具有优化规则的模糊神经网络及其软测量建模

提出了一种通过调整减法聚类半径优选模糊规则的软测量建模方法。首先用减法聚类建立T—S模糊模型，然后通过调整聚类半径优选模糊规则数，以取得具有良好泛化性能的模型，之后利用梯度下降混合最小二乘算法精调参数。最后用该方法对初馏塔石脑油干点进行软测量建模，结果表明能较快确定优化模型，并能满足软测量建模精度要求。

Generating Fuzzy-neural Networks with Optimal Fuzzy Rules Based on Subtractive Clustering with Applications to Soft Sensor Modeling

A soft sensor modeling method is presented which selects optimal fuzzy rules by tuning the radius of a subtractive cluster center. Subtractive clustering is used to generate a T-S fuzzy model. Secondly, the radius of a cluster center is adjusted to select optimal fuzzy rules, to acquire a fuzzy model with perfect generalization capability. The parameters is fine-tuned by means of a hybrid gradient descent (GD) and least-squares estimation (LSE). Finally, the method is used to model a PDU naphtha's dry point and the result shows that it can determine the optimal model fastly and achieve satisfactory prediction precision.