贝叶斯核学习方法在复杂工业系统建模中的应用

Application of Bayesian Kernel Learning in Complex Industrial Process

在复杂工业过程中,需要对过程变量和操作变量进行测量,但是由于客观条件的影响,很难实现直接测量,但是在工业过程中存在大量有用的过程数据,利用这些数据可以挖掘出大量有用的信息。因此可以利用软测量技术建立辅助变量与测量变量的关系。数据模型的建立是软测量技术的关键,贝叶斯核学习方法是一种新型的数据建模方法,具有小样本、推广能力强和收敛速度快的特点。

In complex industrial process, the process variables and operating variables should be measured. But in fact, it is difficult torealize the direct measurement. Those useful data in industrial process can dug up a lot of valuable information. So we can take use thesoft measurement technology to establish relationship between the auxiliary variables and measurement variables. The establishment ofa data model is the key to the soft measurement technology, the Bayesian Kernel Learning method is a kind of new data modelingmethod with the advantage of small samples, the characteristics of strong generalization ability and convergence speed.