A multivariate statistical combination forecasting method for product quality evaluation

In this paper, a multivariate statistical combination forecasting method is proposed for key performance evaluation in process industry. This method is developed based on some of the most popular multivariate statistic approaches. It merges the advantages of the principal component regression method (PCR), the partial least squares regression method (PLSR) and the modified partial least squares regression method (MPLSR). We test the proposed method with a numerical example and also an actual wine production process. The results indicate that the prediction accuracy of the optimal combination forecasting method is superior to those of individual methods.