工业过程模糊控制建模中数据预处理方法研究

Approach for Data Preprocessing on the Fuzzy Control in Industrial Process

工业过程模糊控制中建模是关键,而数据预处理直接影响着模型收敛速度、计算量、模型质量。针对所采集的样本数据提出几个新的数据预处理方法。利用三次样条进行空缺数据的插值填充,然后利用K-均值进行输入输出空间划分与数据离散化,为模糊规则的提取做好准备。

Modeling is the key for fuzzy control in industrial process. And data preprocessing affects the convergence, speed, the calculated value and the quality of the modei directly. In this paper, some new approach for data preprocessing is proposed ac-cording to the collected sample data. A cubic-sp line is applied to interpolate and pack the mull data, and the K-sverage is used to do the input-out put space assignment and the data discrimination. Thus good preparations are made for the extraction of fuzzy rules.