Skip to Main Content
Combining kernel principal component analysis (KPCA) and support vector machines (SVM) in this study, we set up a KPCA-SVM model to predict the fineness of raw mill finished products. We conducted nonlinear feature extraction from the technological parameter samples of the raw mill by means of KPCA and obtained the feature principal components that are easier for regression operations. Thus, the number of input space dimensions that can lower the SVM was met. Then we conducted training by using the least squares support vector machines (LS-SVM). Finally, our computation results proved that the model proposed in this study can effectively predict the fineness of raw mill finished products.