论文部分内容阅读
建立1种基于最小二乘支持向量机(least squares support vector machine,LSSVM)的模糊辨识方法,根据学习样本集的模糊聚类结果,产生LSSVM的模糊核函数,并证明该模糊核函数是Mercer核函数,为LSSVM提供1种构造核函数的简便方法。此外,由于所建立的模糊辨识方法在T-S模糊模型的后件参数学习过程中采用结构风险最小化准则,提高了模型的泛化能力。利用所建立的辨识方法进行热工对象逆系统模型辨识,证明了该方法的有效性。
A fuzzy identification method based on least squares support vector machine (LSSVM) is established. According to the fuzzy clustering result of learning sample set, a fuzzy kernel function of LSSVM is generated and it is proved that the fuzzy kernel function is Mercer kernel Function, providing an easy way for LSSVM to construct kernel functions. In addition, due to the fuzzy identification method established in the T-S fuzzy model’s post-part parameter learning process using structural risk minimization criteria, improve the generalization ability of the model. Using the established identification method to identify the thermal system inverse system model, the effectiveness of the method is proved.