论文部分内容阅读
输入变量个数会对模糊建模精度产生影响。对于一个实际的复杂系统,可测的或者需要考虑的输入变量非常多。是不是考虑的影响因素越多,即模糊系统的输入变量越多,则辨识的效果就越好呢?本文基于T-S模糊模型,分别采用对称三角形模糊划分和网格对角线法以及模糊聚类划分提取模糊规则,对Box-Jenkins煤气炉数据和Mackey-Glass混沌时间序列进行建模,得到了模糊模型训练性能指标和检验性能指标随输入变量个数增加时的变化趋势曲线,并给出了结论。
Enter the number of variables will have an impact on the accuracy of fuzzy modeling. For a real complex system, there are a lot of input variables that can be measured or need to be considered. Whether the more influential factors are considered, that is, the more input variables of the fuzzy system, the better the recognition effect. Based on the TS fuzzy model, this paper uses symmetrical triangular fuzzy partition and grid diagonal method and fuzzy clustering The fuzzy rules are extracted and divided, and the Box-Jenkins gas stove data and Mackey-Glass chaotic time series are modeled. The training performance indexes of fuzzy model and the trend curves of testing performance indexes with the increase of input variables are obtained. in conclusion.