Fast prediction model based big data system identification

In this paper, a fast identification based on ring die granulator system by using prediction model linear LSSVM regression is discussed for big data system. Because the model of regression prediction based on SVM is suitable for small data, the accuracy of regression prediction is not high. However, if the number of data and dimension of feature increase, the training time of model will increase dramatically. In order to solve the problem of long modeling time for inputting large data, the improved NDCD method is used for solving the models. Meanwhile, real data is conducted on the granulator to prove the effect. Compared with other methods for large data system by the simulation, this method has not only apparent advantages but also high fitness. In conclusion, this method has good ability of fast modeling and generation, which can be used real prediction on hoop standard granulator by online prediction model to solve the problem that large time is delayed in outputting of hoop standard granulator.